dtk23.04初始化

megatron/mpu/tests/test_initialize.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from commons import print_separator
+from commons import initialize_distributed
+import mpu
+import torch
+import sys
+sys.path.append("../..")
+
+
+def test_initialize_model_parallel(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing initialize_model_parallel with size {} ...'.format(
+            tensor_model_parallel_size))
+    tensor_model_parallel_size_ = min(tensor_model_parallel_size,
+                               torch.distributed.get_world_size())
+    assert not mpu.model_parallel_is_initialized()
+    mpu.initialize_model_parallel(tensor_model_parallel_size_)
+    assert mpu.model_parallel_is_initialized()
+
+    # Checks.
+    def check(group, world_size, rank):
+        assert world_size == torch.distributed.get_world_size(group=group)
+        assert rank == torch.distributed.get_rank(group=group)
+
+    # Model parallel.
+    world_size = tensor_model_parallel_size_
+    rank = torch.distributed.get_rank() % tensor_model_parallel_size_
+    assert world_size == mpu.get_tensor_model_parallel_world_size()
+    assert rank == mpu.get_tensor_model_parallel_rank()
+    check(mpu.get_tensor_model_parallel_group(), world_size, rank)
+
+    # Data parallel.
+    world_size = torch.distributed.get_world_size() // tensor_model_parallel_size_
+    rank = torch.distributed.get_rank() // tensor_model_parallel_size
+    assert world_size == mpu.get_data_parallel_world_size()
+    assert rank == mpu.get_data_parallel_rank()
+    check(mpu.get_data_parallel_group(), world_size, rank)
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+def test_get_tensor_model_parallel_src_rank(tensor_model_parallel_size_):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing get_tensor_model_parallel_src_rank with size {} ...'.format(
+            tensor_model_parallel_size_))
+    tensor_model_parallel_size = min(tensor_model_parallel_size_,
+                              torch.distributed.get_world_size())
+    assert not mpu.model_parallel_is_initialized()
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    assert mpu.model_parallel_is_initialized()
+
+    # Checks
+    src_rank = torch.distributed.get_rank() - mpu.get_tensor_model_parallel_rank()
+    assert mpu.get_tensor_model_parallel_src_rank() == src_rank
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+if __name__ == '__main__':
+
+    initialize_distributed()
+    world_size = torch.distributed.get_world_size()
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        print_separator('test initialize model parallel')
+        test_initialize_model_parallel(tensor_model_parallel_size)
+        print_separator('test model parallel source rank')
+        test_get_tensor_model_parallel_src_rank(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2

megatron/mpu/tests/test_layers.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from mpu import layers
+from commons import set_random_seed
+from commons import print_separator
+from commons import initialize_distributed
+import mpu
+from torch.nn.parameter import Parameter
+import torch.nn.init as init
+import torch
+import random
+import sys
+sys.path.append("../..")
+
+
+def test_parallel_embedding(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing parallel embedding with model parallel size {} ...'.
+              format(tensor_model_parallel_size))
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    batch_size = 17
+    seq_length = 23
+    vocab_size = 48
+    hidden_size = 16
+    seed = 1236
+
+    set_random_seed(123)
+    input_data = torch.LongTensor(
+        size=(batch_size, seq_length)).random_(0, vocab_size).cuda()
+    loss_weight = torch.randn([batch_size, seq_length, hidden_size]).cuda()
+
+    set_random_seed(seed)
+    embedding_original = torch.nn.Embedding(vocab_size, hidden_size).cuda()
+
+    output = embedding_original(input_data)
+    loss_original = torch.mul(output, loss_weight).sum()
+    loss_original.backward()
+
+    set_random_seed(seed)
+    embedding_parallel = layers.ParallelEmbedding(
+        vocab_size, hidden_size, init_method=init.normal_).cuda()
+    output = embedding_parallel(input_data)
+    loss_parallel = torch.mul(output, loss_weight).sum()
+    loss_parallel.backward()
+
+    set_random_seed(seed)
+    embedding_vocab_parallel = layers.VocabParallelEmbedding(
+        vocab_size, hidden_size, init_method=init.normal_).cuda()
+    output = embedding_vocab_parallel(input_data)
+    loss_vocab_parallel = torch.mul(output, loss_weight).sum()
+    loss_vocab_parallel.backward()
+
+    torch.distributed.barrier()
+    error = loss_parallel.sub(loss_original).abs()
+    print('   error in loss (parallel) on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-12, 'error: {}'.format(error)
+
+    torch.distributed.barrier()
+    error = loss_vocab_parallel.sub(loss_original).abs()
+    print('   error in loss (vocab parallel) on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-12, 'error: {}'.format(error)
+
+    weight_grad_orig = torch.split(embedding_original.weight.grad,
+                                   hidden_size // tensor_model_parallel_size,
+                                   1)[mpu.get_tensor_model_parallel_rank()]
+    error = embedding_parallel.weight.grad.sub(weight_grad_orig).abs().max()
+    print('   error in grad (parallel) on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-12, 'error: {}'.format(error)
+
+    weight_grad_orig = torch.split(embedding_original.weight.grad,
+                                   vocab_size // tensor_model_parallel_size,
+                                   0)[mpu.get_tensor_model_parallel_rank()]
+    error = embedding_vocab_parallel.weight.grad.sub(
+        weight_grad_orig).abs().max()
+    print('   error in grad (vocab parallel) on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-12, 'error: {}'.format(error)
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+def test_initialize_affine_weight(tensor_model_parallel_size):
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    if torch.distributed.get_rank() == 0:
+        print('> testing initialize_affine_weight with model parallel '
+              'size: {}'.format(tensor_model_parallel_size))
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed = 12345
+    input_size_coeff = 13
+    input_size = input_size_coeff * tensor_model_parallel_size
+    output_size_coeff = 17
+    output_size = output_size_coeff * tensor_model_parallel_size
+
+    # ---------------
+    # Column parallel
+    # ---------------
+    weight = torch.empty(output_size_coeff, input_size)
+    set_random_seed(seed)
+    layers._initialize_affine_weight(weight, output_size, input_size,
+
+                                     output_size_coeff, 0,
+                                     torch.nn.init.normal_)
+    # Target.
+    set_random_seed(seed)
+    master_weight = torch.empty(output_size, input_size)
+    torch.nn.init.normal_(master_weight)
+    rank = mpu.get_tensor_model_parallel_rank()
+    my_weight = torch.split(master_weight, output_size_coeff,
+                            dim=0)[rank].contiguous().clone()
+
+    # Compare.
+    error = weight.sub(my_weight).abs().max()
+    torch.distributed.barrier()
+    print('   column parallel max error (should be zero) on global rank '
+          '{}: {}'.format(torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # ------------
+    # Row parallel
+    # ------------
+    weight = torch.empty(output_size, input_size_coeff)
+    set_random_seed(seed)
+    mpu.layers._initialize_affine_weight(weight, output_size, input_size,
+                                         input_size_coeff, 1,
+                                         torch.nn.init.normal_)
+    # Target.
+    set_random_seed(seed)
+    master_weight = torch.empty(output_size, input_size)
+    torch.nn.init.normal_(master_weight)
+    rank = mpu.get_tensor_model_parallel_rank()
+    my_weight = torch.split(master_weight, input_size_coeff,
+                            dim=1)[rank].contiguous().clone()
+
+    # Compare.
+    error = weight.sub(my_weight).abs().max()
+    torch.distributed.barrier()
+    print('   row parallel max error (should be zero) on global rank '
+          '{}: {}'.format(torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print(' >> passed the test :-)')
+
+
+class IdentityLayer2D(torch.nn.Module):
+    def __init__(self, m, n):
+        super(IdentityLayer2D, self).__init__()
+        self.weight = Parameter(torch.Tensor(m, n))
+        torch.nn.init.xavier_normal_(self.weight)
+
+    def forward(self):
+        return self.weight
+
+
+def test_column_parallel_linear(tensor_model_parallel_size):
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    if torch.distributed.get_rank() == 0:
+        print('> testing ColumnParallelLinear with model parallel '
+              'size: {}'.format(tensor_model_parallel_size))
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed = 12345
+    set_random_seed(seed)
+    input_size_coeff = 13
+    input_size = input_size_coeff * tensor_model_parallel_size
+    output_size_coeff = 17
+    output_size = output_size_coeff * tensor_model_parallel_size
+    batch_size = 7
+
+    # Network
+    identity_layer = IdentityLayer2D(batch_size, input_size).cuda()
+    linear_layer = mpu.ColumnParallelLinear(
+        input_size, output_size, keep_master_weight_for_test=True).cuda()
+    loss_weight = torch.randn([batch_size, output_size]).cuda()
+    # Forward
+    input_ = identity_layer()
+    output = linear_layer(input_)
+    loss = torch.mul(output, loss_weight).sum()
+    # Backward
+    loss.backward()
+
+    # Values.
+    dLdY = loss_weight
+    X = identity_layer.weight
+    A = linear_layer.master_weight.cuda()
+    dLdA = torch.matmul(dLdY.t(), X)
+    dLdb = torch.matmul(torch.ones(batch_size, 1).cuda().t(), dLdY).view(-1)
+    dLdX = torch.matmul(dLdY, A)
+
+    rank = mpu.get_tensor_model_parallel_rank()
+    my_dLdA = torch.split(dLdA, output_size_coeff,
+                          dim=0)[rank].contiguous().clone()
+    error = my_dLdA.sub(linear_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdA on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    my_dLdb = torch.split(dLdb, output_size_coeff,
+                          dim=0)[rank].contiguous().clone()
+    error = my_dLdb.sub(linear_layer.bias.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdb on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    error = dLdX.sub(identity_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdX on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print(' >> passed the test :-)')
+
+
+def test_row_parallel_linear(tensor_model_parallel_size):
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    if torch.distributed.get_rank() == 0:
+        print('> testing RowParallelLinear with model parallel '
+              'size: {}'.format(tensor_model_parallel_size))
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed = 12345
+    set_random_seed(seed)
+    input_size_coeff = 13
+    input_size = input_size_coeff * tensor_model_parallel_size
+    output_size_coeff = 17
+    output_size = output_size_coeff * tensor_model_parallel_size
+    batch_size = 7
+
+    # Network
+    identity_layer = IdentityLayer2D(batch_size, input_size).cuda()
+    linear_layer = mpu.RowParallelLinear(
+        input_size, output_size, keep_master_weight_for_test=True).cuda()
+    loss_weight = torch.randn([batch_size, output_size]).cuda()
+    # Forward
+    input_ = identity_layer()
+    output = linear_layer(input_)
+    loss = torch.mul(output, loss_weight).sum()
+    # Backward
+    loss.backward()
+
+    # Values.
+    dLdY = loss_weight
+    X = identity_layer.weight
+    A = linear_layer.master_weight.cuda()
+    dLdA = torch.matmul(dLdY.t(), X)
+    dLdb = torch.matmul(torch.ones(batch_size, 1).cuda().t(), dLdY).view(-1)
+    dLdX = torch.matmul(dLdY, A)
+
+    rank = mpu.get_tensor_model_parallel_rank()
+    my_dLdA = torch.split(dLdA, input_size_coeff,
+                          dim=1)[rank].contiguous().clone()
+    error = my_dLdA.sub(linear_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdA on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    error = dLdb.sub(linear_layer.bias.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdb on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    error = dLdX.sub(identity_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   error in dLdX on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print(' >> passed the test :-)')
+
+
+class IdentityLayer3D(torch.nn.Module):
+    def __init__(self, m, n, k):
+        super(IdentityLayer3D, self).__init__()
+        self.weight = Parameter(torch.Tensor(m, n, k))
+        torch.nn.init.xavier_normal_(self.weight)
+
+    def forward(self):
+        return self.weight
+
+
+def parallel_self_attention(tensor_model_parallel_size, num_att_heads_per_partition,
+                            hidden_size_per_att_head, dropout_prob, batch_size,
+                            sequence_length):
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed = 12345
+    set_random_seed(seed)
+
+    num_att_heads = num_att_heads_per_partition * \
+        torch.distributed.get_world_size()
+    hidden_size = hidden_size_per_att_head * num_att_heads
+
+    # Network
+    identity_layer = IdentityLayer3D(batch_size, sequence_length,
+                                     hidden_size).cuda()
+    attention_layer = mpu.BertParallelSelfAttention(hidden_size, num_att_heads,
+                                                    dropout_prob).cuda()
+    loss_weight = torch.randn([batch_size, sequence_length, hidden_size]).cuda()
+    attention_mask = torch.randn([batch_size, 1, 1, sequence_length]).cuda()
+    # Forward
+    input_ = identity_layer()
+    output = attention_layer(input_, attention_mask)
+    loss = torch.mul(output, loss_weight).sum()
+    # Backward
+    loss.backward()
+
+    rank = mpu.get_tensor_model_parallel_rank()
+    mpu.destroy_model_parallel()
+    return rank, hidden_size, tensor_model_parallel_size, loss, \
+        attention_layer, identity_layer
+
+
+def test_parallel_self_attention(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing ParallelSelfAttention with model parallel '
+              'size: {}'.format(tensor_model_parallel_size))
+
+    num_att_heads_per_partition = 3
+    hidden_size_per_att_head = 7
+    dropout_prob = 0.0  # has to be zero
+    batch_size = 5
+    sequence_length = 13
+
+    rank_1, hideen_size_1, tensor_model_parallel_size_1, loss_1, \
+        attention_layer_1, identity_layer_1 = parallel_self_attention(
+            1, num_att_heads_per_partition,
+            hidden_size_per_att_head, dropout_prob, batch_size, sequence_length)
+
+    rank, hidden_size, tensor_model_parallel_size, loss, \
+        attention_layer, identity_layer = parallel_self_attention(
+            tensor_model_parallel_size, num_att_heads_per_partition,
+            hidden_size_per_att_head, dropout_prob, batch_size, sequence_length)
+    assert hideen_size_1 == hidden_size
+
+    error = loss_1.sub(loss).abs().max()
+    torch.distributed.barrier()
+    print('   loss error on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 5.0e-6
+
+    my_lin_grad_list = torch.split(
+        attention_layer_1.query_key_value.weight.grad,
+        hidden_size // tensor_model_parallel_size, 0)[rank::tensor_model_parallel_size]
+    my_lin_grad = torch.cat(my_lin_grad_list, dim=0)
+    error = my_lin_grad.sub(
+        attention_layer.query_key_value.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   weight gradient error on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 5.0e-6
+
+    error = identity_layer_1.weight.grad.sub(
+        identity_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   input gradient error on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 5.0e-6
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print(' >> passed the test :-)')
+
+
+def parallel_transformer(tensor_model_parallel_size, num_att_heads_per_partition,
+                         hidden_size_per_att_head, batch_size, sequence_length):
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed = 12345
+    set_random_seed(seed)
+
+    num_att_heads = num_att_heads_per_partition * \
+        torch.distributed.get_world_size()
+    hidden_size = hidden_size_per_att_head * num_att_heads
+    intermediate_size = 4 * hidden_size
+
+    # Network
+    identity_layer = IdentityLayer3D(batch_size, sequence_length,
+                                     hidden_size).cuda()
+    transformer_layer = mpu.BertParallelTransformerLayer(
+        hidden_size, intermediate_size, num_att_heads, 0.0, 0.0,
+        torch.nn.functional.relu, 1.0e-5).cuda()
+
+    loss_weight = torch.randn([batch_size, sequence_length, hidden_size]).cuda()
+    attention_mask = torch.randn([batch_size, 1, 1, sequence_length]).cuda()
+    # Forward
+    input_ = identity_layer()
+    output = transformer_layer(input_, attention_mask)
+    loss = torch.mul(output, loss_weight).sum()
+    # Backward
+    loss.backward()
+
+    rank = mpu.get_tensor_model_parallel_rank()
+    mpu.destroy_model_parallel()
+    return rank, hidden_size, tensor_model_parallel_size, loss, \
+        transformer_layer, identity_layer
+
+
+def test_parallel_transformer_layer(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing ParallelTransformerLayer with model parallel '
+              'size: {}'.format(tensor_model_parallel_size))
+
+    num_att_heads_per_partition = 3
+    hidden_size_per_att_head = 7
+    batch_size = 5
+    sequence_length = 13
+
+    rank_1, hidden_size_1, tensor_model_parallel_size_1, loss_1, \
+        transformer_layer_1, identity_layer_1 = parallel_transformer(
+            1, num_att_heads_per_partition,
+            hidden_size_per_att_head, batch_size, sequence_length)
+
+    rank, hidden_size, tensor_model_parallel_size, loss, \
+        transformer_layer, identity_layer = parallel_transformer(
+            tensor_model_parallel_size, num_att_heads_per_partition,
+            hidden_size_per_att_head, batch_size, sequence_length)
+
+    error = loss_1.sub(loss).abs().max()
+    torch.distributed.barrier()
+    print('   loss error on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 5.0e-5, 'error: {}'.format(error)
+
+    error = identity_layer_1.weight.grad.sub(
+        identity_layer.weight.grad).abs().max()
+    torch.distributed.barrier()
+    print('   input gradient error on global rank {}: {}'.format(
+        torch.distributed.get_rank(), error))
+    assert error < 5.0e-5, 'error: {}'.format(error)
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print(' >> passed the test :-)')
+
+
+if __name__ == '__main__':
+
+    torch.backends.cudnn.deterministic = True
+    torch.backends.cudnn.benchmark = False
+
+    initialize_distributed()
+    world_size = torch.distributed.get_world_size()
+
+    print_separator('test initialize affine weight')
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        test_initialize_affine_weight(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        print_separator('test parallel embedding')
+        test_parallel_embedding(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    print_separator('test column-parallel linear')
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        test_column_parallel_linear(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    print_separator('test row-parallel linear')
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        test_row_parallel_linear(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    print_separator('test parallel self-attention')
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        test_parallel_self_attention(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    print_separator('test parallel transformer')
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        test_parallel_transformer_layer(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2

megatron/mpu/tests/test_random.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from commons import print_separator
+from commons import initialize_distributed
+import mpu
+import torch
+import sys
+sys.path.append("../..")
+
+
+def test_set_cuda_rng_state(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing set_rng_state with size {} ...'.
+              format(tensor_model_parallel_size))
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    size = 123
+    seed = 1234
+    torch.cuda.manual_seed(1234)
+    tensor = torch.cuda.FloatTensor(size)
+
+    # Get the state
+    rng_state = torch.cuda.get_rng_state()
+    rng_state_copy = rng_state.clone()
+
+    # Do some stuff.
+    for _ in range(5):
+        torch.randn(size, out=tensor)
+    result_1 = tensor.clone()
+
+    assert rng_state.sub(rng_state_copy).max() == 0
+    assert torch.cuda.get_rng_state().sub(rng_state_copy).max() > 0
+
+    # State should be different.
+    new_rng_state = torch.cuda.get_rng_state()
+    max_diff = new_rng_state.sub(rng_state).max()
+    print('   max diff in rng state (should be non-zero) on global rank {}: {}'.
+          format(torch.distributed.get_rank(), max_diff))
+    assert max_diff > 0
+
+    # Reset the rng state and do the same stuff.
+    mpu.random._set_cuda_rng_state(rng_state)
+    for _ in range(5):
+        torch.randn(size, out=tensor)
+    mpu.random._set_cuda_rng_state(rng_state)
+    for _ in range(5):
+        torch.randn(size, out=tensor)
+    result_2 = tensor.clone()
+
+    # Results should be the same
+    error = result_2.sub(result_1).abs().max()
+    print('   max error in generated tensors (should be zero) on '
+          'global rank {}: {}'.format(torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # Input state should have remained intact.
+    error = rng_state.sub(rng_state_copy).max()
+    print('   max error in rng state (should be zero) on global rank {}: {}'.
+          format(torch.distributed.get_rank(), error))
+    assert error == 0
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+def test_cuda_rng_tracker(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing cuda rng tracker with size {} ...'.
+              format(tensor_model_parallel_size))
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    seed_1 = 1234
+    seed_2 = 4321
+    size = [12, 21]
+    tensor = torch.cuda.FloatTensor(size)
+
+    # Set to seed_1 and generate two tensors.
+    torch.cuda.manual_seed(seed_1)
+    torch.randn(size, out=tensor)
+    target_11 = tensor.clone()
+    torch.randn(size, out=tensor)
+    target_12 = tensor.clone()
+
+    # Set to seed_2 and generate two tensors.
+    torch.cuda.manual_seed(seed_2)
+    torch.randn(size, out=tensor)
+    target_21 = tensor.clone()
+    torch.randn(size, out=tensor)
+    target_22 = tensor.clone()
+
+    # Now if we interleave seed_1 and seed_2,
+    # we should still get the same tensors
+    torch.cuda.manual_seed(seed_1)
+    mpu.get_cuda_rng_tracker().add('test', seed_2)
+
+    torch.randn(size, out=tensor)
+    result_11 = tensor.clone()
+
+    with mpu.get_cuda_rng_tracker().fork('test'):
+        torch.randn(size, out=tensor)
+        result_21 = tensor.clone()
+
+    torch.randn(size, out=tensor)
+    result_12 = tensor.clone()
+
+    with mpu.get_cuda_rng_tracker().fork('test'):
+        torch.randn(size, out=tensor)
+        result_22 = tensor.clone()
+
+    diff = result_11.sub(result_21).abs().max()
+    diff = min(diff, result_12.sub(result_22).abs().max())
+    print('   max diff in generated tensors (should be non-zero) on '
+          'global rank {}: {}'.format(torch.distributed.get_rank(), diff))
+    assert diff > 1.0e-6
+    error = max(result_11.sub(target_11).abs().max(),
+                result_12.sub(target_12).abs().max())
+    error = max(error, result_21.sub(target_21).abs().max())
+    error = max(error, result_22.sub(target_22).abs().max())
+    print('   max error in generated tensors (should be zero) on '
+          'global rank {}: {}'.format(torch.distributed.get_rank(), error))
+    assert error < 1.0e-6
+
+    # Reset the tracker
+    mpu.get_cuda_rng_tracker().reset()
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+def test_model_parallel_cuda_manual_seed(tensor_model_parallel_size):
+
+    if torch.distributed.get_rank() == 0:
+        print('> testing model parallel cuda manual seed with size {} ...'.
+              format(tensor_model_parallel_size))
+
+    mpu.initialize_model_parallel(tensor_model_parallel_size)
+    tensor_model_parallel_size = mpu.get_tensor_model_parallel_world_size()
+
+    mpu.model_parallel_cuda_manual_seed(12345)
+    assert torch.cuda.initial_seed() == 12345
+    with mpu.get_cuda_rng_tracker().fork():
+        assert torch.cuda.initial_seed() == (12345 + 2718 +
+                                             mpu.get_tensor_model_parallel_rank())
+
+    # Reset the tracker
+    mpu.get_cuda_rng_tracker().reset()
+
+    # Reset groups
+    mpu.destroy_model_parallel()
+
+    torch.distributed.barrier()
+    if torch.distributed.get_rank() == 0:
+        print('>> passed the test :-)')
+
+
+if __name__ == '__main__':
+
+    initialize_distributed()
+    world_size = torch.distributed.get_world_size()
+
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        print_separator('test set rng state')
+        test_set_cuda_rng_state(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        print_separator('test cuda rng tracker')
+        test_cuda_rng_tracker(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2
+
+    tensor_model_parallel_size = 1
+    while tensor_model_parallel_size <= world_size:
+        print_separator('test model parallel cuda manual seed')
+        test_model_parallel_cuda_manual_seed(tensor_model_parallel_size)
+        tensor_model_parallel_size *= 2

megatron/mpu/utils.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+import torch
+
+
+def ensure_divisibility(numerator, denominator):
+    """Ensure that numerator is divisible by the denominator."""
+    assert numerator % denominator == 0, '{} is not divisible by {}'.format(
+        numerator, denominator)
+
+
+def divide(numerator, denominator):
+    """Ensure that numerator is divisible by the denominator and return
+    the division value."""
+    ensure_divisibility(numerator, denominator)
+    return numerator // denominator
+
+
+def split_tensor_along_last_dim(tensor, num_partitions,
+                                contiguous_split_chunks=False):
+    """Split a tensor along its last dimension.
+    Arguments:
+        tensor: input tensor.
+        num_partitions: number of partitions to split the tensor
+        contiguous_split_chunks: If True, make each chunk contiguous
+                                 in memory.
+    """
+    # Get the size and dimension.
+    last_dim = tensor.dim() - 1
+    last_dim_size = divide(tensor.size()[last_dim], num_partitions)
+    # Split.
+    tensor_list = torch.split(tensor, last_dim_size, dim=last_dim)
+    # Note: torch.split does not create contiguous tensors by default.
+    if contiguous_split_chunks:
+        return tuple(chunk.contiguous() for chunk in tensor_list)
+
+    return tensor_list
+
+
+class VocabUtility:
+    """Split the vocabulary into `world_size` chunks amd return the
+        first and last index of the vocabulary belonging to the `rank`
+        partition: Note that indecies in [fist, last)"""
+
+    @staticmethod
+    def vocab_range_from_per_partition_vocab_size(per_partition_vocab_size,
+                                                  rank, world_size):
+        index_f = rank * per_partition_vocab_size
+        index_l = index_f + per_partition_vocab_size
+        return index_f, index_l
+
+    @staticmethod
+    def vocab_range_from_global_vocab_size(global_vocab_size, rank, world_size):
+        per_partition_vocab_size = divide(global_vocab_size, world_size)
+        return VocabUtility.vocab_range_from_per_partition_vocab_size(
+            per_partition_vocab_size, rank, world_size)

megatron/optimizer/__init__.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from apex.optimizers import FusedAdam as Adam
+from apex.optimizers import FusedSGD as SGD
+
+from megatron import get_args
+from megatron.model.fused_layer_norm import MixedFusedLayerNorm as LayerNorm
+
+from .grad_scaler import ConstantGradScaler, DynamicGradScaler
+from .optimizer import Float16OptimizerWithFloat16Params, FP32Optimizer
+
+
+def _get_params_for_weight_decay_optimization(modules):
+    """Divide params into with-weight-decay and without-weight-decay groups.
+    Layernorms and baises will have no weight decay but the rest will.
+    """
+
+    weight_decay_params = {'params': []}
+    no_weight_decay_params = {'params': [], 'weight_decay': 0.0}
+    for module in modules:
+        for module_ in module.modules():
+            if isinstance(module_, LayerNorm):
+                no_weight_decay_params['params'].extend(
+                    [p for p in list(module_._parameters.values())
+                     if p is not None])
+            else:
+                weight_decay_params['params'].extend(
+                    [p for n, p in list(module_._parameters.items())
+                     if p is not None and n != 'bias'])
+                no_weight_decay_params['params'].extend(
+                    [p for n, p in list(module_._parameters.items())
+                     if p is not None and n == 'bias'])
+
+    # XXX: temp hack to workaround the crash in apex FusedAdam's multi_tensor_applier
+    #
+    # it crashes when the param count is larger than a certain size which we hit at 200B over 80
+    # A100 gpus - I think around 2.7B per gpu, so halving it works around the issue
+    param_count = len(weight_decay_params['params'])
+    first_half = weight_decay_params['params'][:param_count // 2]
+    second_half = weight_decay_params['params'][param_count // 2:]
+
+    first_half =  { 'params': first_half }
+    second_half = { 'params': second_half }
+
+    return first_half, second_half, no_weight_decay_params
+
+    #return weight_decay_params, no_weight_decay_params
+
+
+def get_megatron_optimizer(model):
+    args = get_args()
+
+    if args.cpu_optimizer:
+        raise NotImplementedError('need to add cpu adam')
+
+    # Base optimizer.
+    param_groups = _get_params_for_weight_decay_optimization(model)
+    if args.optimizer == 'adam':
+        if args.use_bnb_optimizer:
+            import bitsandbytes as bnb
+            adam_optimizer = bnb.optim.Adam8bit
+        else:
+            adam_optimizer = Adam
+        optimizer = adam_optimizer(param_groups,
+                                   lr=args.lr,
+                                   weight_decay=args.weight_decay,
+                                   betas=(args.adam_beta1, args.adam_beta2),
+                                   eps=args.adam_eps)
+    elif args.optimizer == 'sgd':
+        optimizer = SGD(param_groups,
+                        lr=args.lr,
+                        weight_decay=args.weight_decay,
+                        momentum=args.sgd_momentum)
+    else:
+        raise Exception('{} optimizer is not supported.'.format(
+            args.optimizer))
+
+    if args.deepspeed:
+        return optimizer
+
+    # Determine whether the params have main-grad field.
+    params_have_main_grad = False
+    if args.DDP_impl == 'local':
+        params_have_main_grad = True
+
+    if args.fp16 or args.bf16:
+
+        # Grad scaler:
+        #    if loss-scale is provided, instantiate the constant scaler.
+        #    if we are using fp16 and loss-scale is not present, use a
+        #       dynamic scaler.
+        #    otherwise we are running in bf16 with no loss-scale so
+        #       leave it as None.
+        grad_scaler = None
+        # Constant loss scale.
+        if args.loss_scale:
+            grad_scaler = ConstantGradScaler(args.loss_scale)
+        # Dynamic loss scale.
+        else:
+            if args.fp16:
+                grad_scaler = DynamicGradScaler(
+                    initial_scale=args.initial_loss_scale,
+                    min_scale=args.min_loss_scale,
+                    growth_factor=2.0,
+                    backoff_factor=0.5,
+                    growth_interval=args.loss_scale_window,
+                    hysteresis=args.hysteresis)
+
+        # Megatron optimizer.
+        return Float16OptimizerWithFloat16Params(optimizer,
+                                                 args.clip_grad,
+                                                 args.log_num_zeros_in_grad,
+                                                 params_have_main_grad,
+                                                 args.bf16,
+                                                 grad_scaler)
+
+    # FP32.
+    return FP32Optimizer(optimizer, args.clip_grad,
+                         args.log_num_zeros_in_grad,
+                         params_have_main_grad)

megatron/optimizer/clip_grads.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Gradient clipping."""
+
+import torch
+from torch._six import inf
+
+from apex.multi_tensor_apply import multi_tensor_applier
+import amp_C
+
+from megatron import mpu
+from megatron.model.module import param_is_not_shared
+from megatron.mpu.layers import param_is_not_tensor_parallel_duplicate
+
+
+def clip_grad_norm_fp32(parameters, max_norm, norm_type=2):
+    """Clips gradient norm of an iterable of parameters whose gradients
+       are in fp32.
+
+    This is adapted from torch.nn.utils.clip_grad.clip_grad_norm_ and
+    added functionality to handle model parallel parameters. Note that
+    the gradients are modified in place.
+
+    Arguments:
+        parameters (Iterable[Tensor] or Tensor): an iterable of Tensors or a
+            single Tensor that will have gradients normalized
+        max_norm (float or int): max norm of the gradients
+        norm_type (float or int): type of the used p-norm. Can be ``'inf'`` for
+            infinity norm.
+
+    Returns:
+        Total norm of the parameters (viewed as a single vector).
+    """
+
+    if isinstance(parameters, torch.Tensor):
+        parameters = [parameters]
+
+    # Filter parameters based on:
+    #   - grad should not be none
+    #   - parameter should not be shared
+    #   - should not be a replica due to tensor model parallelism
+    grads = []
+    grads_for_norm = []
+    for param in parameters:
+        grad_not_none = param.grad is not None
+        is_not_shared = param_is_not_shared(param)
+        is_not_tp_duplicate = param_is_not_tensor_parallel_duplicate(param)
+        grad = param.grad.detach()
+        if grad_not_none:
+            # Make sure the grads are in fp32
+            assert param.grad.type() == 'torch.cuda.FloatTensor'
+            grads.append(grad)
+        if grad_not_none and is_not_shared and is_not_tp_duplicate:
+            grads_for_norm.append(grad)
+
+    # Norm parameters.
+    max_norm = float(max_norm)
+    norm_type = float(norm_type)
+    total_norm = 0.0
+
+    # Calculate norm.
+    if norm_type == inf:
+        total_norm = max(grad.abs().max() for grad in grads_for_norm)
+        total_norm_cuda = torch.cuda.FloatTensor([float(total_norm)])
+        # Take max across all model-parallel GPUs.
+        torch.distributed.all_reduce(total_norm_cuda,
+                                     op=torch.distributed.ReduceOp.MAX,
+                                     group=mpu.get_model_parallel_group())
+        total_norm = total_norm_cuda[0].item()
+
+    else:
+        if norm_type == 2.0:
+            dummy_overflow_buf = torch.cuda.IntTensor([0])
+            # Use apex's multi-tensor applier for efficiency reasons.
+            # Multi-tensor applier takes a function and a list of list
+            # and performs the operation on that list all in one kernel.
+            grad_norm, _ = multi_tensor_applier(
+                amp_C.multi_tensor_l2norm,
+                dummy_overflow_buf,
+                [grads_for_norm],
+                False # no per-parameter norm
+            )
+            # Since we will be summing across data parallel groups,
+            # we need the pow(norm-type).
+            total_norm = grad_norm ** norm_type
+
+        else:
+            for grad in grads_for_norm:
+                grad_norm = torch.norm(grad, norm_type)
+                total_norm += grad_norm ** norm_type
+
+        # Sum across all model-parallel GPUs.
+        torch.distributed.all_reduce(total_norm,
+                                     op=torch.distributed.ReduceOp.SUM,
+                                     group=mpu.get_model_parallel_group())
+        total_norm = total_norm.item() ** (1.0 / norm_type)
+
+    # Scale.
+    clip_coeff = max_norm / (total_norm + 1.0e-6)
+    if clip_coeff < 1.0:
+        dummy_overflow_buf = torch.cuda.IntTensor([0])
+        multi_tensor_applier(amp_C.multi_tensor_scale,
+                             dummy_overflow_buf,
+                             [grads, grads],
+                             clip_coeff)
+
+    return total_norm
+
+
+def count_zeros_fp32(parameters):
+
+    if isinstance(parameters, torch.Tensor):
+        parameters = [parameters]
+
+    # Filter parameters based on:
+    #   - grad should not be none
+    #   - parameter should not be shared
+    #   - should not be a replica due to tensor model parallelism
+    total_num_zeros = 0.0
+    for param in parameters:
+        grad_not_none = param.grad is not None
+        is_not_shared = param_is_not_shared(param)
+        is_not_tp_duplicate = param_is_not_tensor_parallel_duplicate(param)
+        if grad_not_none and is_not_shared and is_not_tp_duplicate:
+            grad = param.grad.detach()
+            num_zeros = grad.numel() - torch.count_nonzero(grad)
+            total_num_zeros = num_zeros + total_num_zeros
+
+    # Sum across all model-parallel GPUs.
+    torch.distributed.all_reduce(total_num_zeros,
+                                 op=torch.distributed.ReduceOp.SUM,
+                                 group=mpu.get_model_parallel_group())
+    total_num_zeros = total_num_zeros.item()
+
+    return total_num_zeros

megatron/optimizer/grad_scaler.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Megatron grad scaler."""
+
+from abc import ABC
+from abc import abstractmethod
+
+import torch
+
+
+class MegatronGradScaler(ABC):
+
+    def __init__(self, initial_scale):
+        """Initialize scale value with the input initial scale."""
+        assert initial_scale > 0.0
+        self._scale = torch.cuda.FloatTensor([initial_scale])
+
+    @property
+    def scale(self):
+        return self._scale
+
+    @property
+    def inv_scale(self):
+        return self._scale.double().reciprocal().float()
+
+    @abstractmethod
+    def update(self, found_inf):
+        pass
+
+    @abstractmethod
+    def state_dict(self):
+        pass
+
+    @abstractmethod
+    def load_state_dict(self, state_dict):
+        pass
+
+
+
+class ConstantGradScaler(MegatronGradScaler):
+
+    def update(self, found_inf):
+        pass
+
+    def state_dict(self):
+        return dict()
+
+    def load_state_dict(self, state_dict):
+        pass
+
+
+
+class DynamicGradScaler(MegatronGradScaler):
+
+    def __init__(self, initial_scale, min_scale,
+                 growth_factor, backoff_factor,
+                 growth_interval, hysteresis):
+        """"Grad scaler with dynamic scale that gets adjusted
+        during training."""
+        super(DynamicGradScaler, self).__init__(initial_scale)
+
+        # Lower bound on the scale.
+        assert min_scale > 0.0
+        assert min_scale <= initial_scale
+        self.min_scale = torch.cuda.FloatTensor([min_scale])
+        # Growth and backoff factors for the scale.
+        assert growth_factor > 1.0
+        self.growth_factor = torch.cuda.FloatTensor([growth_factor])
+        assert backoff_factor < 1.0
+        assert backoff_factor > 0.0
+        self.backoff_factor = torch.cuda.FloatTensor([backoff_factor])
+        # Interval over which if we don't see any inf/nan,
+        # we will scale the grad scale by the growth factor.
+        assert growth_interval > 0
+        self.growth_interval = growth_interval
+        # Number of inf/nans we should see before scaling down
+        # the grad scale by the backoff factor.
+        assert hysteresis > 0
+        self.hysteresis = hysteresis
+
+        # Trackers.
+        self._growth_tracker = 0
+        self._hysteresis_tracker = self.hysteresis
+
+
+    def update(self, found_inf):
+
+        # If we have an inf/nan, growth tracker is set to 0
+        # and hysterisis tracker is reduced by 1.
+        if found_inf:
+            self._growth_tracker = 0
+            self._hysteresis_tracker -= 1
+            # Now if we are out of hysteresis count, scale down the loss.
+            if self._hysteresis_tracker <= 0:
+                self._scale = torch.max(self._scale * self.backoff_factor,
+                                        self.min_scale)
+        else:
+            # If there is no nan/inf, increment the growth tracker.
+            self._growth_tracker += 1
+            # If we have had enough consequitive intervals with no nan/inf:
+            if self._growth_tracker == self.growth_interval:
+                # Reset the tracker and hysteresis trackers,
+                self._growth_tracker = 0
+                self._hysteresis_tracker = self.hysteresis
+                # and scale up the loss scale.
+                self._scale = self._scale * self.growth_factor
+
+
+    def state_dict(self):
+        state_dict = {}
+        state_dict['scale'] = self._scale
+        state_dict['growth_tracker'] = self._growth_tracker
+        state_dict['hysteresis_tracker'] = self._hysteresis_tracker
+        return state_dict
+
+
+    def load_state_dict(self, state_dict):
+        self._scale = state_dict['scale'].cuda(torch.cuda.current_device())
+        self._growth_tracker = state_dict['growth_tracker']
+        self._hysteresis_tracker = state_dict['hysteresis_tracker']

megatron/optimizer/optimizer.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Megatron optimizer."""
+
+from abc import ABC
+from abc import abstractmethod
+
+import torch
+
+from apex.multi_tensor_apply import multi_tensor_applier
+import amp_C
+
+from megatron import get_timers
+from megatron import mpu
+from megatron import print_rank_0
+
+from .clip_grads import clip_grad_norm_fp32, count_zeros_fp32
+
+
+def _zero_grad_group_helper(group, set_to_none):
+    """Zero out the gradient for a group of parameters.
+    Note: copied from torch.optim.optimizer."""
+    for param in group:
+        if param.grad is not None:
+            if set_to_none:
+                param.grad = None
+            else:
+                if param.grad.grad_fn is not None:
+                    param.grad.detach_()
+                else:
+                    param.grad.requires_grad_(False)
+                param.grad.zero_()
+
+
+def _multi_tensor_copy_this_to_that(this, that, overflow_buf=None):
+    """Use multi-tensor-applier to copy values from one list to another.
+    We don't have a blfoat16 implementation so for now if the overflow_buf
+    is not provided, we default back to simple loop copy to be compatible
+    with bfloat16."""
+    if overflow_buf:
+        overflow_buf.fill_(0)
+        # Scaling with factor `1.0` is equivalent to copy.
+        multi_tensor_applier(amp_C.multi_tensor_scale,
+                             overflow_buf,
+                             [this, that],
+                             1.0)
+    else:
+        for this_, that_ in zip(this, that):
+            that_.copy_(this_)
+
+
+
+class MegatronOptimizer(ABC):
+
+
+    def __init__(self, optimizer, clip_grad,
+                 log_num_zeros_in_grad,
+                 params_have_main_grad):
+        """Input optimizer is the base optimizer for example Adam."""
+        self.optimizer = optimizer
+        assert self.optimizer, 'no optimizer is provided.'
+        # Set gradient clipping and logging params.
+        self.clip_grad = clip_grad
+        self.log_num_zeros_in_grad = log_num_zeros_in_grad
+        self.params_have_main_grad = params_have_main_grad
+
+
+    def get_parameters(self):
+        params = []
+        for param_group in self.optimizer.param_groups:
+            for param in param_group['params']:
+                params.append(param)
+        return params
+
+
+    def clip_grad_norm(self, clip_grad):
+        params = self.get_parameters()
+        return clip_grad_norm_fp32(params, clip_grad)
+
+
+    def count_zeros(self):
+        params = self.get_parameters()
+        return count_zeros_fp32(params)
+
+
+    @abstractmethod
+    def zero_grad(self, set_to_none=True):
+        pass
+
+
+    @abstractmethod
+    def get_loss_scale(self):
+        """The output should be a cuda tensor of size 1."""
+        pass
+
+
+    def scale_loss(self, loss):
+        """Simple scaling."""
+        return self.get_loss_scale() * loss
+
+
+    @abstractmethod
+    def step(self):
+        pass
+
+
+    @abstractmethod
+    def reload_model_params(self):
+        """Refreshes any internal state from the current model parameters.
+        Call whenever the parameters are changed outside of the optimizer.
+        For example, when we load a model from a checkpoint  without loading
+        the optimizer, the model parameters are updated but for fp16 optimizer
+        with main parameters, the main parameters need to also be updated."""
+        pass
+
+
+    @abstractmethod
+    def state_dict(self):
+        pass
+
+
+    @abstractmethod
+    def load_state_dict(self, state_dict):
+        pass
+
+
+    # Promote state so it can be retrieved or set via
+    # "optimizer_instance.state"
+    def _get_state(self):
+        return self.optimizer.state
+
+    def _set_state(self, value):
+        self.optimizer.state = value
+
+    state = property(_get_state, _set_state)
+
+
+    # Promote param_groups so it can be retrieved or set via
+    # "optimizer_instance.param_groups"
+    # (for example, to adjust the learning rate)
+    def _get_param_groups(self):
+        return self.optimizer.param_groups
+
+    def _set_param_groups(self, value):
+        self.optimizer.param_groups = value
+
+    param_groups = property(_get_param_groups, _set_param_groups)
+
+
+
+class Float16OptimizerWithFloat16Params(MegatronOptimizer):
+    """Float16 optimizer for fp16 and bf16 data types.
+
+    Arguments:
+        optimizer: base optimizer such as Adam or SGD
+        clip_grad: clip gradeints with this global L2 norm. Note
+            that clipping is ignored if clip_grad == 0
+        log_num_zeros_in_grad: return number of zeros in the gradients.
+        params_have_main_grad: flag indicating if parameters have
+            a `main_grad` field. If this is set, we are assuming
+            that the model parameters are store in the `main_grad`
+            field instead of the typical `grad` field. This happens
+            for the DDP cases where there is a contihuous buffer
+            holding the gradients. For example for bfloat16, we want
+            to do gradient accumulation and all-reduces in float32
+            and as a result we store those gradients in the main_grad.
+            Note that main grad is not necessarily in float32.
+        bf16: if true, the model is running in bfloat16.
+        grad_scaler: used for scaling gradients. Note that this can be
+            None. This case happens when `bf16 = True` and we don't
+            use any loss scale. Note that for `bf16 = True`, we can have
+            a constnat gradient scaler. Also for `bf16 = False`, we
+            always require a grad scaler.
+    """
+
+    def __init__(self, optimizer, clip_grad, log_num_zeros_in_grad,
+                 params_have_main_grad, bf16, grad_scaler):
+
+        super(Float16OptimizerWithFloat16Params, self).__init__(
+            optimizer, clip_grad, log_num_zeros_in_grad,
+            params_have_main_grad)
+
+        self.bf16 = bf16
+        self.grad_scaler = grad_scaler
+        # None grad scaler is only supported for bf16.
+        if self.grad_scaler is None:
+            assert self.bf16, 'fp16 expects a grad scaler.'
+
+        # Tensor used to determine if a nan/if has happend.
+        # Any non-zero value indicates inf/nan.
+        # Note that we keep this for the cases that grad scaler is none.
+        # We still record nan/inf if we have a bfloat16 with a grad scaler.
+        if self.grad_scaler:
+            self.found_inf = torch.cuda.FloatTensor([0.0])
+
+        # Dummy tensor needed for apex multi-apply tensor.
+        # For bfloat, we don't have multi-tensor apply and for now
+        # we set it to none so the multi-tensor apply gets ignored.
+        if bf16:
+            self._dummy_overflow_buf = None
+        else:
+            self._dummy_overflow_buf = torch.cuda.IntTensor([0])
+
+        # In case grad scaler is not passed, define the unity scale.
+        if self.grad_scaler is None:
+            self._scale_one = torch.cuda.FloatTensor([1.0])
+
+        # ======================
+        # main parameter stuff
+        # ======================
+
+        # Three groups of parameters:
+        #   float16_groups: original float16 parameters
+        #   fp32_from_float16_groups: fp32 copy of float16 parameters
+        #   fp32_from_fp32_groups: original fp32 parameters
+        self.float16_groups = []
+        self.fp32_from_float16_groups = []
+        self.fp32_from_fp32_groups = []
+
+        # For all the groups in the original optimizer:
+        for param_group in self.optimizer.param_groups:
+            float16_params_this_group = []
+            fp32_params_this_group = []
+            fp32_from_float16_params_this_group = []
+            # For all the parameters in this group:
+            for i, param in enumerate(param_group['params']):
+                if param.requires_grad:
+
+                    # float16 params:
+                    if param.type() in ['torch.cuda.HalfTensor',
+                                        'torch.cuda.BFloat16Tensor']:
+                        float16_params_this_group.append(param)
+                        # Create a copy
+                        main_param = param.detach().clone().float()
+                        # Copy tensor model parallel attributes.
+                        mpu.copy_tensor_model_parallel_attributes(main_param,
+                                                                  param)
+                        if hasattr(param, 'shared'):
+                            main_param.shared = param.shared
+                        # Replace the optimizer params with the new fp32 copy.
+                        param_group['params'][i] = main_param
+                        fp32_from_float16_params_this_group.append(main_param)
+                        # Reset existing state dict key to the new main param.
+                        if param in self.optimizer.state:
+                            self.optimizer.state[main_param] \
+                                = self.optimizer.state.pop(param)
+
+                    # fp32 params.
+                    elif param.type() == 'torch.cuda.FloatTensor':
+                        fp32_params_this_group.append(param)
+                        param_group['params'][i] = param
+
+                    else:
+                        raise TypeError('Wrapped parameters must be one of '
+                                        'torch.cuda.FloatTensor,  '
+                                        'torch.cuda.HalfTensor, or '
+                                        'torch.cuda.BFloat16Tensor. '
+                                        'Received {}'.format(param.type()))
+
+            self.float16_groups.append(float16_params_this_group)
+            self.fp32_from_float16_groups.append(
+                fp32_from_float16_params_this_group)
+            self.fp32_from_fp32_groups.append(fp32_params_this_group)
+
+        # Leverage state_dict() and load_state_dict() to
+        # recast preexisting per-param state tensors
+        self.optimizer.load_state_dict(self.optimizer.state_dict())
+
+
+    def zero_grad(self, set_to_none=True):
+        """We only need to zero the model related parameters, i.e.,
+                float16_groups & fp32_from_fp32_groups."""
+        for group in self.float16_groups:
+            _zero_grad_group_helper(group, set_to_none)
+        for group in self.fp32_from_fp32_groups:
+            _zero_grad_group_helper(group, set_to_none)
+
+
+    def get_loss_scale(self):
+        if self.grad_scaler is None:
+            return self._scale_one
+        return self.grad_scaler.scale
+
+
+    def _copy_model_grads_to_main_grads(self):
+        # This only needs to be done for the float16 group.
+        for model_group, main_group in zip(self.float16_groups,
+                                           self.fp32_from_float16_groups):
+            for model_param, main_param in zip(model_group, main_group):
+                if self.params_have_main_grad:
+                    main_param.grad = model_param.main_grad.float()
+                else:
+                    if model_param.grad is not None:
+                        main_param.grad = model_param.grad.float()
+        # For fp32 grads, we need to reset the grads to main grad.
+        if self.params_have_main_grad:
+            for model_group in self.fp32_from_fp32_groups:
+                for model_param in model_group:
+                    model_param.grad = model_param.main_grad
+
+
+    def _unscale_main_grads_and_check_for_nan(self):
+        main_grads = []
+        # fp32 params fromm float16 ones.
+        for main_group in self.fp32_from_float16_groups:
+            for main_param in main_group:
+                if main_param.grad is not None:
+                    main_grads.append(main_param.grad.data)
+        # Append fp32 parameters.
+        for main_group in self.fp32_from_fp32_groups:
+            for main_param in main_group:
+                if main_param.grad is not None:
+                    main_grads.append(main_param.grad.data)
+        # Reset found inf.
+        self.found_inf.fill_(0.0)
+        # Unscale and set found inf/nan
+        torch._amp_foreach_non_finite_check_and_unscale_(
+            main_grads, self.found_inf, self.grad_scaler.inv_scale)
+        # Update across all model parallel instances.
+        torch.distributed.all_reduce(self.found_inf,
+                                     op=torch.distributed.ReduceOp.MAX,
+                                     group=mpu.get_model_parallel_group())
+
+        # Check for nan.
+        found_inf_flag = (self.found_inf.item() > 0)
+        return found_inf_flag
+
+
+    def _get_model_and_main_params_data_float16(self):
+        model_data = []
+        main_data = []
+        for model_group, main_group in zip(self.float16_groups,
+                                           self.fp32_from_float16_groups):
+            for model_param, main_param in zip(model_group, main_group):
+                model_data.append(model_param.data)
+                main_data.append(main_param.data)
+        return model_data, main_data
+
+
+    def _copy_main_params_to_model_params(self):
+        # Only needed for the float16 params.
+        model_data, main_data = self._get_model_and_main_params_data_float16()
+        _multi_tensor_copy_this_to_that(this=main_data, that=model_data,
+                                        overflow_buf=self._dummy_overflow_buf)
+
+
+    def _copy_model_params_to_main_params(self):
+        # Only needed for the float16 params.
+        model_data, main_data = self._get_model_and_main_params_data_float16()
+        _multi_tensor_copy_this_to_that(this=model_data, that=main_data,
+                                        overflow_buf=self._dummy_overflow_buf)
+
+
+    def reload_model_params(self):
+        self._copy_model_params_to_main_params()
+
+
+    @torch.no_grad()
+    def step(self):
+
+        timers = get_timers()
+
+        # Copy gradients from model params to main params.
+        timers('optimizer-copy-to-main-grad').start()
+        self._copy_model_grads_to_main_grads()
+        timers('optimizer-copy-to-main-grad').stop()
+
+        # Do unscale, check for inf, and update grad scaler only for
+        # the case that grad scaler is provided.
+        if self.grad_scaler:
+
+            # Unscale and check for inf/nan.
+            timers('optimizer-unscale-and-check-inf').start()
+            found_inf_flag = self._unscale_main_grads_and_check_for_nan()
+            timers('optimizer-unscale-and-check-inf').stop()
+
+            # We are done with scaling gradients
+            # so we can update the loss scale.
+            self.grad_scaler.update(found_inf_flag)
+
+            # If we found inf/nan, skip the update.
+            if found_inf_flag:
+                return False, None, None
+
+        # Clip the main gradients.
+        timers('optimizer-clip-main-grad').start()
+        grad_norm = None
+        if self.clip_grad > 0.0:
+            grad_norm = self.clip_grad_norm(self.clip_grad)
+        timers('optimizer-clip-main-grad').stop()
+
+        # count the zeros in the grads
+        num_zeros_in_grad = self.count_zeros() if \
+                            self.log_num_zeros_in_grad else None
+
+        # Step the optimizer.
+        self.optimizer.step()
+
+        # Update params from main params.
+        timers('optimizer-copy-main-to-model-params').start()
+        self._copy_main_params_to_model_params()
+        timers('optimizer-copy-main-to-model-params').stop()
+
+        # Successful update.
+        return True, grad_norm, num_zeros_in_grad
+
+
+    def state_dict(self):
+        state_dict = {}
+        state_dict['optimizer'] = self.optimizer.state_dict()
+        if self.grad_scaler:
+            state_dict['grad_scaler'] = self.grad_scaler.state_dict()
+        state_dict['fp32_from_fp16_params'] = self.fp32_from_float16_groups
+        return state_dict
+
+
+    def load_state_dict(self, state_dict):
+        # Optimizer.
+        optimizer_key = 'optimizer'
+        if optimizer_key not in state_dict:
+            optimizer_key = 'optimizer_state_dict'
+            print_rank_0('***WARNING*** loading optimizer from '
+                         'an old checkpoint ...')
+        self.optimizer.load_state_dict(state_dict[optimizer_key])
+
+        # Grad scaler.
+        if 'grad_scaler' not in state_dict:
+            print_rank_0('***WARNING*** found an old checkpoint, will not '
+                         'load grad scaler ...')
+        else:
+            if self.grad_scaler:
+                self.grad_scaler.load_state_dict(state_dict['grad_scaler'])
+            else:
+                print_rank_0('***WARNING*** fould the grad scaler in the '
+                             'checkpoint but it is None in the class. '
+                             'Skipping loading grad scaler ...')
+
+        # Copy data for the main params.
+        fp32_from_float16_params_key = 'fp32_from_fp16_params'
+        if fp32_from_float16_params_key not in state_dict:
+            fp32_from_float16_params_key = 'fp32_from_fp16'
+        for current_group, saved_group in zip(
+                self.fp32_from_float16_groups,
+                state_dict[fp32_from_float16_params_key]):
+            for current_param, saved_param in zip(current_group, saved_group):
+                current_param.data.copy_(saved_param.data)
+
+
+
+class FP32Optimizer(MegatronOptimizer):
+
+    def __init__(self, optimizer, clip_grad,
+                 log_num_zeros_in_grad,
+                 params_have_main_grad):
+
+        super(FP32Optimizer, self).__init__(
+            optimizer, clip_grad, log_num_zeros_in_grad,
+            params_have_main_grad)
+
+        self._scale = torch.cuda.FloatTensor([1.0])
+
+
+    def zero_grad(self, set_to_none=True):
+        """Copied from torch.optim.optimizer"""
+        for group in self.optimizer.param_groups:
+            _zero_grad_group_helper(group['params'], set_to_none)
+
+
+    def get_loss_scale(self):
+        """FP32 optimizer does not do any scaling."""
+        return self._scale
+
+
+    @torch.no_grad()
+    def step(self):
+        """Clip gradients (if needed) and step the base optimizer.
+        Always return successful since there is no overflow."""
+
+        # Copy main_grads to grads.
+        if self.params_have_main_grad:
+            for param_group in self.optimizer.param_groups:
+                for param in param_group['params']:
+                    param.grad = param.main_grad
+
+        # Clip gradients.
+        grad_norm = None
+        if self.clip_grad > 0.0:
+            grad_norm = self.clip_grad_norm(self.clip_grad)
+
+        # count the zeros in the grads
+        num_zeros_in_grad = self.count_zeros() if \
+                            self.log_num_zeros_in_grad else None
+
+        # Update parameters.
+        self.optimizer.step()
+
+        # No overflow for FP32 optimizer.
+        return True, grad_norm, num_zeros_in_grad
+
+
+    def reload_model_params(self):
+        pass
+
+
+    def state_dict(self):
+        return self.optimizer.state_dict()
+
+
+    def load_state_dict(self, state_dict):
+        self.optimizer.load_state_dict(state_dict)

megatron/p2p_communication.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from functools import reduce
+import operator
+import torch
+
+from megatron import get_args
+from megatron import mpu
+
+
+def _communicate(tensor_send_next, tensor_send_prev, recv_prev, recv_next,
+                 use_ring_exchange=False):
+    """Communicate tensors between stages. Used as helper method in other
+    communication methods that are used in megatron/schedules.py.
+
+    Takes the following arguments:
+        tensor_send_next: tensor to send to next rank (no tensor sent if
+                          set to None).
+        tensor_send_prev: tensor to send to prev rank (no tensor sent if
+                          set to None).
+        recv_prev: boolean for whether tensor should be received from
+                   previous rank.
+        recv_next: boolean for whether tensor should be received from
+                   next rank.
+        use_ring_exchange: boolean for whether torch.distributed.ring_exchange()
+                           API should be used.
+
+    Returns:
+        (tensor_recv_prev, tensor_recv_next)
+    """
+    args = get_args()
+
+    # Create placeholder tensors for receive in forward and backward directions
+    # if needed.
+    tensor_recv_prev = None
+    tensor_recv_next = None
+    tensor_shape = (args.seq_length, args.micro_batch_size, args.hidden_size)
+    if args.scatter_gather_tensors_in_pipeline:
+        tensor_chunk_shape = reduce(operator.mul, tensor_shape, 1) // \
+            mpu.get_tensor_model_parallel_world_size()
+    else:
+        tensor_chunk_shape = tensor_shape
+    dtype = args.params_dtype
+    if args.fp32_residual_connection:
+        dtype = torch.float
+    if recv_prev:
+        tensor_recv_prev = torch.empty(tensor_chunk_shape,
+                                       requires_grad=True,
+                                       device=torch.cuda.current_device(),
+                                       dtype=dtype)
+    if recv_next:
+        tensor_recv_next = torch.empty(tensor_chunk_shape,
+                                       requires_grad=True,
+                                       device=torch.cuda.current_device(),
+                                       dtype=dtype)
+
+    # Split tensor into smaller chunks if using scatter-gather optimization.
+    if args.scatter_gather_tensors_in_pipeline:
+        if tensor_send_next is not None:
+            tensor_send_next = mpu.split_tensor_into_1d_equal_chunks(tensor_send_next)
+
+        if tensor_send_prev is not None:
+            tensor_send_prev = mpu.split_tensor_into_1d_equal_chunks(tensor_send_prev)
+
+    # Send tensors in both the forward and backward directions as appropriate.
+    if use_ring_exchange:
+        torch.distributed.ring_exchange(tensor_send_prev=tensor_send_prev,
+                                        tensor_recv_prev=tensor_recv_prev,
+                                        tensor_send_next=tensor_send_next,
+                                        tensor_recv_next=tensor_recv_next,
+                                        group=mpu.get_pipeline_model_parallel_group())
+    else:
+        ops = []
+        if tensor_send_prev is not None:
+            send_prev_op = torch.distributed.P2POp(
+                torch.distributed.isend, tensor_send_prev,
+                mpu.get_pipeline_model_parallel_prev_rank())
+            ops.append(send_prev_op)
+        if tensor_recv_prev is not None:
+            recv_prev_op = torch.distributed.P2POp(
+                torch.distributed.irecv, tensor_recv_prev,
+                mpu.get_pipeline_model_parallel_prev_rank())
+            ops.append(recv_prev_op)
+        if tensor_send_next is not None:
+            send_next_op = torch.distributed.P2POp(
+                torch.distributed.isend, tensor_send_next,
+                mpu.get_pipeline_model_parallel_next_rank())
+            ops.append(send_next_op)
+        if tensor_recv_next is not None:
+            recv_next_op = torch.distributed.P2POp(
+                torch.distributed.irecv, tensor_recv_next,
+                mpu.get_pipeline_model_parallel_next_rank())
+            ops.append(recv_next_op)
+        if len(ops) > 0:
+            reqs = torch.distributed.batch_isend_irecv(ops)
+            for req in reqs:
+                req.wait()
+    # To protect against race condition when using batch_isend_irecv().
+    torch.cuda.synchronize()
+
+    # If using scatter-gather optimization, gather smaller chunks.
+    if args.scatter_gather_tensors_in_pipeline:
+        if recv_prev:
+            tensor_recv_prev = mpu.gather_split_1d_tensor(
+                tensor_recv_prev).view(tensor_shape).requires_grad_()
+
+        if recv_next:
+            tensor_recv_next = mpu.gather_split_1d_tensor(
+                tensor_recv_next).view(tensor_shape).requires_grad_()
+
+    return tensor_recv_prev, tensor_recv_next
+
+
+def recv_forward(timers=None):
+    """Receive tensor from previous rank in pipeline (forward receive)."""
+    if mpu.is_pipeline_first_stage():
+        input_tensor = None
+    else:
+        if timers is not None:
+            timers('forward-recv').start()
+        input_tensor, _ = _communicate(
+            tensor_send_next=None,
+            tensor_send_prev=None,
+            recv_prev=True,
+            recv_next=False)
+        if timers is not None:
+            timers('forward-recv').stop()
+    return input_tensor
+
+
+def recv_backward(timers=None):
+    """Receive tensor from next rank in pipeline (backward receive)."""
+    if mpu.is_pipeline_last_stage():
+        output_tensor_grad = None
+    else:
+        if timers is not None:
+            timers('backward-recv').start()
+        _, output_tensor_grad = _communicate(
+            tensor_send_next=None,
+            tensor_send_prev=None,
+            recv_prev=False,
+            recv_next=True)
+        if timers is not None:
+            timers('backward-recv').stop()
+    return output_tensor_grad
+
+
+def send_forward(output_tensor, timers=None):
+    """Send tensor to next rank in pipeline (forward send)."""
+    if not mpu.is_pipeline_last_stage():
+        if timers is not None:
+            timers('forward-send').start()
+        _communicate(
+            tensor_send_next=output_tensor,
+            tensor_send_prev=None,
+            recv_prev=False,
+            recv_next=False)
+        if timers is not None:
+            timers('forward-send').stop()
+
+
+def send_backward(input_tensor_grad, timers=None):
+    """Send tensor to previous rank in pipeline (backward send)."""
+    if not mpu.is_pipeline_first_stage():
+        if timers is not None:
+            timers('backward-send').start()
+        _communicate(
+            tensor_send_next=None,
+            tensor_send_prev=input_tensor_grad,
+            recv_prev=False,
+            recv_next=False)
+        if timers is not None:
+            timers('backward-send').stop()
+
+
+def send_forward_recv_backward(output_tensor, timers=None):
+    """Batched send and recv with next rank in pipeline."""
+    if mpu.is_pipeline_last_stage():
+        output_tensor_grad = None
+    else:
+        if timers is not None:
+            timers('forward-send-backward-recv').start()
+        _, output_tensor_grad = _communicate(
+            tensor_send_next=output_tensor,
+            tensor_send_prev=None,
+            recv_prev=False,
+            recv_next=True)
+        if timers is not None:
+            timers('forward-send-backward-recv').stop()
+    return output_tensor_grad
+
+
+def send_backward_recv_forward(input_tensor_grad, timers=None):
+    """Batched send and recv with previous rank in pipeline."""
+    if mpu.is_pipeline_first_stage():
+        input_tensor = None
+    else:
+        if timers is not None:
+            timers('backward-send-forward-recv').start()
+        input_tensor, _ = _communicate(
+            tensor_send_next=None,
+            tensor_send_prev=input_tensor_grad,
+            recv_prev=True,
+            recv_next=False)
+        if timers is not None:
+            timers('backward-send-forward-recv').stop()
+    return input_tensor
+
+
+def send_forward_recv_forward(output_tensor, recv_prev, timers=None):
+    """Batched recv from previous rank and send to next rank in pipeline."""
+    if timers is not None:
+        timers('forward-send-forward-recv').start()
+    input_tensor, _ = _communicate(
+        tensor_send_next=output_tensor,
+        tensor_send_prev=None,
+        recv_prev=recv_prev,
+        recv_next=False)
+    if timers is not None:
+        timers('forward-send-forward-recv').stop()
+    return input_tensor
+
+
+def send_backward_recv_backward(input_tensor_grad, recv_next, timers=None):
+    """Batched recv from next rank and send to previous rank in pipeline."""
+    if timers is not None:
+        timers('backward-send-backward-recv').start()
+    _, output_tensor_grad = _communicate(
+        tensor_send_next=None,
+        tensor_send_prev=input_tensor_grad,
+        recv_prev=False,
+        recv_next=recv_next)
+    if timers is not None:
+        timers('backward-send-backward-recv').stop()
+    return output_tensor_grad
+
+
+def send_forward_backward_recv_forward_backward(
+        output_tensor, input_tensor_grad, recv_prev,
+        recv_next, timers=None):
+    """Batched send and recv with previous and next ranks in pipeline."""
+    if timers is not None:
+        timers('forward-backward-send-forward-backward-recv').start()
+    input_tensor, output_tensor_grad = _communicate(
+        tensor_send_next=output_tensor,
+        tensor_send_prev=input_tensor_grad,
+        recv_prev=recv_prev,
+        recv_next=recv_next)
+    if timers is not None:
+        timers('forward-backward-send-forward-backward-recv').stop()
+    return input_tensor, output_tensor_grad

megatron/package_info.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+MAJOR = 1
+MINOR = 1.5
+
+# Use the following formatting: (major, minor)
+VERSION = (MAJOR, MINOR)
+
+__version__ = '.'.join(map(str, VERSION)) + '.bs'
+__package_name__ = 'megatron-lm'
+__contact_names__ = 'NVIDIA INC'
+__url__ = 'https://github.com/NVIDIA/Megatron-LM'
+__download_url__ = 'https://github.com/NVIDIA/Megatron-LM/releases'
+__description__ = 'Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism.'
+__license__ = 'See https://github.com/NVIDIA/Megatron-LM/blob/master/LICENSE'
+__keywords__ = 'deep learning, Megatron, gpu, NLP, nvidia, pytorch, torch, language'

megatron/schedules.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from contextlib import contextmanager
+import torch
+from torch.nn.parallel.distributed import DistributedDataParallel as torchDDP
+
+from megatron import get_args
+from megatron import get_num_microbatches
+from megatron import get_timers
+from megatron import mpu
+from megatron import p2p_communication
+from megatron.utils import unwrap_model
+from megatron.model.distributed import DistributedDataParallel as LocalDDP
+from megatron.model.module import Float16Module
+
+def get_forward_backward_func():
+    args = get_args()
+    if mpu.get_pipeline_model_parallel_world_size() > 1:
+        if args.virtual_pipeline_model_parallel_size is not None:
+            forward_backward_func = forward_backward_pipelining_with_interleaving
+        else:
+            forward_backward_func = forward_backward_pipelining_without_interleaving
+    else:
+        forward_backward_func = forward_backward_no_pipelining
+    return forward_backward_func
+
+
+def forward_step(forward_step_func, data_iterator, model, input_tensor, losses_reduced):
+    """Forward step for passed-in model.
+
+    If first stage, input tensor is obtained from data_iterator, otherwise
+    passed-in input_tensor is used.
+
+    Returns output tensor."""
+    timers = get_timers()
+
+    args = get_args()
+
+    timers('forward-compute').start()
+    unwrapped_model = unwrap_model(
+        model, (torchDDP, LocalDDP, Float16Module))
+    if not args.deepspeed:
+        unwrapped_model.set_input_tensor(input_tensor)
+    else:
+        unwrapped_model.module.set_input_tensor(input_tensor)
+
+    output_tensor, loss_func = forward_step_func(data_iterator, model)
+    if mpu.is_pipeline_last_stage():
+        output_tensor = loss_func(output_tensor)
+        loss, loss_reduced = output_tensor
+        output_tensor = loss / get_num_microbatches()
+        losses_reduced.append(loss_reduced)
+    timers('forward-compute').stop()
+
+    return output_tensor
+
+
+def backward_step(optimizer, input_tensor, output_tensor, output_tensor_grad, model=None):
+    """Backward step through passed-in output tensor.
+
+    If last stage, output_tensor_grad is None, otherwise gradient of loss
+    with respect to stage's output tensor.
+
+    Returns gradient of loss with respect to input tensor (None if first
+    stage)."""
+    args = get_args()
+
+    if args.deepspeed:
+        assert model is not None
+
+    timers = get_timers()
+    timers('backward-compute').start()
+
+    # Retain the grad on the input_tensor.
+    if input_tensor is not None:
+        input_tensor.retain_grad()
+
+    if args.deepspeed:
+        model.backward(output_tensor)
+    else:
+        # Backward pass.
+        if output_tensor_grad is None:
+            output_tensor = optimizer.scale_loss(output_tensor)
+        torch.autograd.backward(output_tensor, grad_tensors=output_tensor_grad)
+
+    # Collect the grad of the input_tensor.
+    input_tensor_grad = None
+    if input_tensor is not None:
+        input_tensor_grad = input_tensor.grad
+
+    timers('backward-compute').stop()
+
+    return input_tensor_grad
+
+
+@contextmanager
+def dummy_handler():
+    try:
+        yield
+    finally:
+        pass
+
+
+def forward_backward_no_pipelining(forward_step_func, data_iterator, model,
+                                   optimizer, timers, forward_only):
+    """Run forward and backward passes with no pipeline parallelism
+    (no inter-stage communication).
+
+    Returns dictionary with losses."""
+    assert len(model) == 1
+    model = model[0]
+
+    context_handler = dummy_handler
+    if isinstance(model, torchDDP):
+        context_handler = model.no_sync
+
+    losses_reduced = []
+    input_tensor, output_tensor_grad = None, None
+    with context_handler():
+        for i in range(get_num_microbatches() - 1):
+            output_tensor = forward_step(forward_step_func, data_iterator, model,
+                                         input_tensor, losses_reduced)
+            if not forward_only:
+                backward_step(optimizer, input_tensor, output_tensor,
+                              output_tensor_grad, model)
+
+    # Run computation for last microbatch out of context handler (want to
+    # synchronize gradients).
+    output_tensor = forward_step(forward_step_func, data_iterator, model,
+                                 input_tensor, losses_reduced)
+    if not forward_only:
+        backward_step(optimizer, input_tensor, output_tensor, output_tensor_grad, model)
+
+    return losses_reduced
+
+
+def forward_backward_pipelining_with_interleaving(forward_step_func, data_iterator, model,
+                                                  optimizer, timers, forward_only):
+    """Run interleaved 1F1B schedule (model split into model chunks), with
+    communication between pipeline stages as needed.
+
+    Returns dictionary with losses if the last stage, empty dict otherwise."""
+    input_tensors = [[] for _ in range(len(model))]
+    output_tensors = [[] for _ in range(len(model))]
+    losses_reduced = []
+    if not forward_only:
+        output_tensor_grads = [[] for _ in range(len(model))]
+
+    pipeline_parallel_size = mpu.get_pipeline_model_parallel_world_size()
+    pipeline_parallel_rank = mpu.get_pipeline_model_parallel_rank()
+
+    # Compute number of warmup and remaining microbatches.
+    num_model_chunks = len(model)
+    num_microbatches = get_num_microbatches() * num_model_chunks
+    all_warmup_microbatches = False
+    if forward_only:
+        num_warmup_microbatches = num_microbatches
+    else:
+        # Run all forward passes and then all backward passes if number of
+        # microbatches is just the number of pipeline stages.
+        # Otherwise, perform (num_model_chunks-1)*pipeline_parallel_size on
+        # all workers, followed by more microbatches after depending on
+        # stage ID (more forward passes for earlier stages, later stages can
+        # immediately start with 1F1B).
+        if get_num_microbatches() == pipeline_parallel_size:
+            num_warmup_microbatches = num_microbatches
+            all_warmup_microbatches = True
+        else:
+            num_warmup_microbatches = \
+                (pipeline_parallel_size - pipeline_parallel_rank - 1) * 2
+            num_warmup_microbatches += (
+                num_model_chunks - 1) * pipeline_parallel_size
+            num_warmup_microbatches = min(num_warmup_microbatches,
+                                          num_microbatches)
+    num_microbatches_remaining = \
+        num_microbatches - num_warmup_microbatches
+
+    def get_model_chunk_id(microbatch_id, forward):
+        """Helper method to get the model chunk ID given the iteration number."""
+        microbatch_id_in_group = microbatch_id % (pipeline_parallel_size * num_model_chunks)
+        model_chunk_id = microbatch_id_in_group // pipeline_parallel_size
+        if not forward:
+            model_chunk_id = (num_model_chunks - model_chunk_id - 1)
+        return model_chunk_id
+
+    def forward_step_helper(microbatch_id):
+        """Helper method to run forward step with model split into chunks
+        (run set_virtual_pipeline_model_parallel_rank() before calling
+        forward_step())."""
+        model_chunk_id = get_model_chunk_id(microbatch_id, forward=True)
+        mpu.set_virtual_pipeline_model_parallel_rank(model_chunk_id)
+
+        if mpu.is_pipeline_first_stage():
+            if len(input_tensors[model_chunk_id]) == \
+                    len(output_tensors[model_chunk_id]):
+                input_tensors[model_chunk_id].append(None)
+        input_tensor = input_tensors[model_chunk_id][-1]
+        output_tensor = forward_step(forward_step_func,
+                                     data_iterator[model_chunk_id],
+                                     model[model_chunk_id],
+                                     input_tensor, losses_reduced)
+        output_tensors[model_chunk_id].append(output_tensor)
+
+        return output_tensor
+
+    def backward_step_helper(microbatch_id):
+        """Helper method to run backward step with model split into chunks
+        (run set_virtual_pipeline_model_parallel_rank() before calling
+        backward_step())."""
+        model_chunk_id = get_model_chunk_id(microbatch_id, forward=False)
+        mpu.set_virtual_pipeline_model_parallel_rank(model_chunk_id)
+
+        if mpu.is_pipeline_last_stage():
+            if len(output_tensor_grads[model_chunk_id]) == 0:
+                output_tensor_grads[model_chunk_id].append(None)
+        input_tensor = input_tensors[model_chunk_id].pop(0)
+        output_tensor = output_tensors[model_chunk_id].pop(0)
+        output_tensor_grad = output_tensor_grads[model_chunk_id].pop(0)
+        input_tensor_grad = \
+            backward_step(optimizer,
+                          input_tensor,
+                          output_tensor,
+                          output_tensor_grad)
+
+        return input_tensor_grad
+
+    # Run warmup forward passes.
+    mpu.set_virtual_pipeline_model_parallel_rank(0)
+    input_tensors[0].append(
+        p2p_communication.recv_forward(timers))
+    for k in range(num_warmup_microbatches):
+        output_tensor = forward_step_helper(k)
+
+        # Determine if tensor should be received from previous stage.
+        next_forward_model_chunk_id = get_model_chunk_id(k+1, forward=True)
+        recv_prev = True
+        if mpu.is_pipeline_first_stage(ignore_virtual=True):
+            if next_forward_model_chunk_id == 0:
+                recv_prev = False
+        if k == (num_microbatches - 1):
+            recv_prev = False
+
+        # Don't send tensor downstream if on last stage.
+        if mpu.is_pipeline_last_stage():
+            output_tensor = None
+
+        # Send and receive tensors as appropriate (send tensors computed
+        # in this iteration; receive tensors for next iteration).
+        if k == (num_warmup_microbatches - 1) and not forward_only and \
+                not all_warmup_microbatches:
+            input_tensor_grad = None
+            recv_next = True
+            if mpu.is_pipeline_last_stage(ignore_virtual=True):
+                recv_next = False
+            input_tensor, output_tensor_grad = \
+                p2p_communication.send_forward_backward_recv_forward_backward(
+                        output_tensor, input_tensor_grad,
+                        recv_prev=recv_prev, recv_next=recv_next,
+                        timers=timers)
+            output_tensor_grads[num_model_chunks-1].append(output_tensor_grad)
+        else:
+            input_tensor = \
+                p2p_communication.send_forward_recv_forward(
+                    output_tensor, recv_prev, timers)
+        input_tensors[next_forward_model_chunk_id].append(input_tensor)
+
+    # Run 1F1B in steady state.
+    for k in range(num_microbatches_remaining):
+        # Forward pass.
+        forward_k = k + num_warmup_microbatches
+        output_tensor = forward_step_helper(forward_k)
+
+        # Backward pass.
+        backward_k = k
+        input_tensor_grad = backward_step_helper(backward_k)
+
+        # Send output_tensor and input_tensor_grad, receive input_tensor
+        # and output_tensor_grad.
+
+        # Determine if current stage has anything to send in either direction,
+        # otherwise set tensor to None.
+        forward_model_chunk_id = get_model_chunk_id(forward_k, forward=True)
+        mpu.set_virtual_pipeline_model_parallel_rank(forward_model_chunk_id)
+        if mpu.is_pipeline_last_stage():
+            output_tensor = None
+
+        backward_model_chunk_id = get_model_chunk_id(backward_k, forward=False)
+        mpu.set_virtual_pipeline_model_parallel_rank(backward_model_chunk_id)
+        if mpu.is_pipeline_first_stage():
+            input_tensor_grad = None
+
+        # Determine if peers are sending, and where in data structure to put
+        # received tensors.
+        recv_prev = True
+        if mpu.is_pipeline_first_stage(ignore_virtual=True):
+            # First stage is ahead of last stage by (pipeline_parallel_size - 1).
+            next_forward_model_chunk_id = get_model_chunk_id(
+                forward_k - (pipeline_parallel_size - 1), forward=True)
+            if next_forward_model_chunk_id == (num_model_chunks - 1):
+                recv_prev = False
+            next_forward_model_chunk_id += 1
+        else:
+            next_forward_model_chunk_id = get_model_chunk_id(forward_k + 1,
+                                                             forward=True)
+
+        recv_next = True
+        if mpu.is_pipeline_last_stage(ignore_virtual=True):
+            # Last stage is ahead of first stage by (pipeline_parallel_size - 1).
+            next_backward_model_chunk_id = get_model_chunk_id(
+                backward_k - (pipeline_parallel_size - 1), forward=False)
+            if next_backward_model_chunk_id == 0:
+                recv_next = False
+            next_backward_model_chunk_id -= 1
+        else:
+            next_backward_model_chunk_id = get_model_chunk_id(backward_k + 1,
+                                                              forward=False)
+
+        # If last iteration, don't receive; we already received one extra
+        # before the start of the for loop.
+        if k == (num_microbatches_remaining - 1):
+            recv_prev = False
+
+        # Communicate tensors.
+        input_tensor, output_tensor_grad = \
+            p2p_communication.send_forward_backward_recv_forward_backward(
+                    output_tensor, input_tensor_grad,
+                    recv_prev=recv_prev, recv_next=recv_next,
+                    timers=timers)
+
+        # Put input_tensor and output_tensor_grad in data structures in the
+        # right location.
+        if recv_prev:
+            input_tensors[next_forward_model_chunk_id].append(input_tensor)
+        if recv_next:
+            output_tensor_grads[next_backward_model_chunk_id].append(
+                output_tensor_grad)
+
+    # Run cooldown backward passes (flush out pipeline).
+    if not forward_only:
+        if all_warmup_microbatches:
+            output_tensor_grads[num_model_chunks-1].append(
+                p2p_communication.recv_backward(timers))
+        for k in range(num_microbatches_remaining, num_microbatches):
+            input_tensor_grad = backward_step_helper(k)
+            next_backward_model_chunk_id = get_model_chunk_id(k+1, forward=False)
+            recv_next = True
+            if mpu.is_pipeline_last_stage(ignore_virtual=True):
+                if next_backward_model_chunk_id == (num_model_chunks - 1):
+                    recv_next = False
+            if k == (num_microbatches - 1):
+                recv_next = False
+            output_tensor_grads[next_backward_model_chunk_id].append(
+                p2p_communication.send_backward_recv_backward(
+                    input_tensor_grad, recv_next, timers))
+
+    return losses_reduced
+
+
+def forward_backward_pipelining_without_interleaving(forward_step_func, data_iterator,
+                                                     model, optimizer, timers,
+                                                     forward_only):
+    """Run non-interleaved 1F1B schedule, with communication between pipeline
+    stages.
+
+    Returns dictionary with losses if the last stage, empty dict otherwise."""
+    timers = get_timers()
+
+    assert len(model) == 1
+    model = model[0]
+
+    # Compute number of warmup microbatches.
+    num_microbatches = get_num_microbatches()
+    num_warmup_microbatches = \
+        (mpu.get_pipeline_model_parallel_world_size() -
+         mpu.get_pipeline_model_parallel_rank() - 1)
+    num_warmup_microbatches = min(
+        num_warmup_microbatches,
+        num_microbatches)
+    num_microbatches_remaining = \
+        num_microbatches - num_warmup_microbatches
+
+    input_tensors = []
+    output_tensors = []
+    losses_reduced = []
+
+    # Run warmup forward passes.
+    for i in range(num_warmup_microbatches):
+        input_tensor = p2p_communication.recv_forward(timers)
+        output_tensor = forward_step(forward_step_func, data_iterator, model,
+                                     input_tensor, losses_reduced)
+        p2p_communication.send_forward(output_tensor, timers)
+
+        input_tensors.append(input_tensor)
+        output_tensors.append(output_tensor)
+
+    # Before running 1F1B, need to receive first forward tensor.
+    # If all microbatches are run in warmup / cooldown phase, then no need to
+    # receive this tensor here.
+    if num_microbatches_remaining > 0:
+        input_tensor = p2p_communication.recv_forward(timers)
+
+    # Run 1F1B in steady state.
+    for i in range(num_microbatches_remaining):
+        last_iteration = (i == (num_microbatches_remaining - 1))
+
+        output_tensor = forward_step(forward_step_func, data_iterator, model,
+                                     input_tensor, losses_reduced)
+        if forward_only:
+            p2p_communication.send_forward(output_tensor, timers)
+        else:
+            output_tensor_grad = \
+                p2p_communication.send_forward_recv_backward(output_tensor,
+                                                             timers)
+
+        # Add input_tensor and output_tensor to end of list, then pop from the
+        # start of the list for backward pass.
+        input_tensors.append(input_tensor)
+        output_tensors.append(output_tensor)
+
+        if forward_only:
+            if not last_iteration:
+                input_tensor = p2p_communication.recv_forward(timers)
+        else:
+            input_tensor, output_tensor = input_tensors.pop(0), output_tensors.pop(0)
+
+            input_tensor_grad = \
+                backward_step(optimizer, input_tensor, output_tensor,
+                              output_tensor_grad, model)
+
+            if last_iteration:
+                input_tensor = None
+                p2p_communication.send_backward(input_tensor_grad, timers)
+            else:
+                input_tensor = \
+                    p2p_communication.send_backward_recv_forward(
+                        input_tensor_grad, timers)
+
+    # Run cooldown backward passes.
+    if not forward_only:
+        for i in range(num_warmup_microbatches):
+            input_tensor = input_tensors.pop(0)
+            output_tensor = output_tensors.pop(0)
+
+            output_tensor_grad = p2p_communication.recv_backward(timers)
+
+            input_tensor_grad = \
+                backward_step(optimizer, input_tensor, output_tensor,
+                              output_tensor_grad, model)
+
+            p2p_communication.send_backward(input_tensor_grad, timers)
+
+    return losses_reduced

megatron/testing_utils.py

+# Copyright 2020 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import contextlib
+import importlib.util
+import inspect
+import logging
+import numpy as np
+import os
+import random
+import re
+import shutil
+import sys
+import tempfile
+import unittest
+
+from distutils.util import strtobool
+from io import StringIO
+from packaging import version
+from pathlib import Path
+from typing import Iterator, Union
+from unittest import mock
+from unittest.case import SkipTest
+
+
+try:
+    import torch
+    _torch_available = True
+except:
+    _torch_available = False
+
+def is_torch_available():
+    return _torch_available
+
+def parse_flag_from_env(key, default=False):
+    try:
+        value = os.environ[key]
+    except KeyError:
+        # KEY isn't set, default to `default`.
+        _value = default
+    else:
+        # KEY is set, convert it to True or False.
+        try:
+            _value = strtobool(value)
+        except ValueError:
+            # More values are supported, but let's keep the message simple.
+            raise ValueError(f"If set, {key} must be yes or no.")
+    return _value
+
+
+def parse_int_from_env(key, default=None):
+    try:
+        value = os.environ[key]
+    except KeyError:
+        _value = default
+    else:
+        try:
+            _value = int(value)
+        except ValueError:
+            raise ValueError(f"If set, {key} must be a int.")
+    return _value
+
+
+def require_torch(test_case):
+    """
+    Decorator marking a test that requires PyTorch.
+
+    These tests are skipped when PyTorch isn't installed.
+
+    """
+    if not is_torch_available():
+        return unittest.skip("test requires PyTorch")(test_case)
+    else:
+        return test_case
+
+
+
+def require_torch_multi_gpu(test_case):
+    """
+    Decorator marking a test that requires a multi-GPU setup (in PyTorch). These tests are skipped on a machine without
+    multiple GPUs.
+
+    To run *only* the multi_gpu tests, assuming all test names contain multi_gpu: $ pytest -sv ./tests -k "multi_gpu"
+    """
+    if not is_torch_available():
+        return unittest.skip("test requires PyTorch")(test_case)
+
+    import torch
+
+    if torch.cuda.device_count() < 2:
+        return unittest.skip("test requires multiple GPUs")(test_case)
+    else:
+        return test_case
+
+
+def require_torch_non_multi_gpu(test_case):
+    """
+    Decorator marking a test that requires 0 or 1 GPU setup (in PyTorch).
+    """
+    if not is_torch_available():
+        return unittest.skip("test requires PyTorch")(test_case)
+
+    import torch
+
+    if torch.cuda.device_count() > 1:
+        return unittest.skip("test requires 0 or 1 GPU")(test_case)
+    else:
+        return test_case
+
+
+def require_torch_up_to_2_gpus(test_case):
+    """
+    Decorator marking a test that requires 0 or 1 or 2 GPU setup (in PyTorch).
+    """
+    if not is_torch_available():
+        return unittest.skip("test requires PyTorch")(test_case)
+
+    import torch
+
+    if torch.cuda.device_count() > 2:
+        return unittest.skip("test requires 0 or 1 or 2 GPUs")(test_case)
+    else:
+        return test_case
+
+
+def require_torch_tpu(test_case):
+    """
+    Decorator marking a test that requires a TPU (in PyTorch).
+    """
+    if not is_torch_tpu_available():
+        return unittest.skip("test requires PyTorch TPU")
+    else:
+        return test_case
+
+
+if is_torch_available():
+    # Set env var CUDA_VISIBLE_DEVICES="" to force cpu-mode
+    import torch
+
+    torch_device = "cuda" if torch.cuda.is_available() else "cpu"
+else:
+    torch_device = None
+
+
+def require_torch_gpu(test_case):
+    """Decorator marking a test that requires CUDA and PyTorch."""
+    if torch_device != "cuda":
+        return unittest.skip("test requires CUDA")(test_case)
+    else:
+        return test_case
+
+
+def require_datasets(test_case):
+    """Decorator marking a test that requires datasets."""
+
+    if not is_datasets_available():
+        return unittest.skip("test requires `datasets`")(test_case)
+    else:
+        return test_case
+
+def is_deepspeed_available():
+    return importlib.util.find_spec("deepspeed") is not None
+
+def require_deepspeed(test_case):
+    """
+    Decorator marking a test that requires deepspeed
+    """
+    if not is_deepspeed_available():
+        return unittest.skip("test requires deepspeed")(test_case)
+    else:
+        return test_case
+
+def is_bnb_available():
+    return importlib.util.find_spec("bitsandbytes") is not None
+
+def require_bnb(test_case):
+    """
+    Decorator marking a test that requires bitsandbytes
+    """
+    if not is_bnb_available():
+        return unittest.skip("test requires bitsandbytes from https://github.com/facebookresearch/bitsandbytes")(test_case)
+    else:
+        return test_case
+
+
+def require_bnb_non_decorator():
+    """
+    Non-Decorator function that would skip a test if bitsandbytes is missing
+    """
+    if not is_bnb_available():
+        raise SkipTest("Test requires bitsandbytes from https://github.com/facebookresearch/bitsandbytes")
+
+
+def set_seed(seed: int=42):
+    """
+    Helper function for reproducible behavior to set the seed in ``random``, ``numpy``, ``torch``
+
+    Args:
+        seed (:obj:`int`): The seed to set.
+    """
+    random.seed(seed)
+    np.random.seed(seed)
+    if is_torch_available():
+        torch.manual_seed(seed)
+        torch.cuda.manual_seed_all(seed)
+        # ^^ safe to call this function even if cuda is not available
+
+def get_gpu_count():
+    """
+    Return the number of available gpus (regardless of whether torch or tf is used)
+    """
+    if is_torch_available():
+        import torch
+
+        return torch.cuda.device_count()
+    elif is_tf_available():
+        import tensorflow as tf
+
+        return len(tf.config.list_physical_devices("GPU"))
+    else:
+        return 0
+
+def torch_assert_equal(actual, expected, **kwargs):
+    # assert_close was added around pt-1.9, it does better checks - e.g will check dimensions match
+    if hasattr(torch.testing, "assert_close"):
+        return torch.testing.assert_close(actual, expected, rtol=0.0, atol=0.0, **kwargs)
+    else:
+        return torch.allclose(actual, expected, rtol=0.0, atol=0.0)
+
+def torch_assert_close(actual, expected, **kwargs):
+    # assert_close was added around pt-1.9, it does better checks - e.g will check dimensions match
+    if hasattr(torch.testing, "assert_close"):
+        return torch.testing.assert_close(actual, expected, **kwargs)
+    else:
+        kwargs.pop("msg", None) # doesn't have msg arg
+        return torch.allclose(actual, expected, **kwargs)
+
+
+def is_torch_bf16_available():
+    # from https://github.com/huggingface/transformers/blob/26eb566e43148c80d0ea098c76c3d128c0281c16/src/transformers/file_utils.py#L301
+    if is_torch_available():
+        import torch
+        if not torch.cuda.is_available() or torch.version.cuda is None:
+            return False
+        if torch.cuda.get_device_properties(torch.cuda.current_device()).major < 8:
+            return False
+        if int(torch.version.cuda.split(".")[0]) < 11:
+            return False
+        if not version.parse(torch.__version__) >= version.parse("1.09"):
+            return False
+        return True
+    else:
+        return False
+
+
+def require_torch_bf16(test_case):
+    """Decorator marking a test that requires CUDA hardware supporting bf16 and PyTorch >= 1.9."""
+    if not is_torch_bf16_available():
+        return unittest.skip("test requires CUDA hardware supporting bf16 and PyTorch >= 1.9")(test_case)
+    else:
+        return test_case
+
+
+def get_tests_dir(append_path=None):
+    """
+    Args:
+        append_path: optional path to append to the tests dir path
+
+    Return:
+        The full path to the `tests` dir, so that the tests can be invoked from anywhere. Optionally `append_path` is
+        joined after the `tests` dir the former is provided.
+
+    """
+    # this function caller's __file__
+    caller__file__ = inspect.stack()[1][1]
+    tests_dir = os.path.abspath(os.path.dirname(caller__file__))
+    if append_path:
+        return os.path.join(tests_dir, append_path)
+    else:
+        return tests_dir
+
+
+#
+# Helper functions for dealing with testing text outputs
+# The original code came from:
+# https://github.com/fastai/fastai/blob/master/tests/utils/text.py
+
+# When any function contains print() calls that get overwritten, like progress bars,
+# a special care needs to be applied, since under pytest -s captured output (capsys
+# or contextlib.redirect_stdout) contains any temporary printed strings, followed by
+# \r's. This helper function ensures that the buffer will contain the same output
+# with and without -s in pytest, by turning:
+# foo bar\r tar mar\r final message
+# into:
+# final message
+# it can handle a single string or a multiline buffer
+def apply_print_resets(buf):
+    return re.sub(r"^.*\r", "", buf, 0, re.M)
+
+
+def assert_screenout(out, what):
+    out_pr = apply_print_resets(out).lower()
+    match_str = out_pr.find(what.lower())
+    assert match_str != -1, f"expecting to find {what} in output: f{out_pr}"
+
+
+class CaptureStd:
+    """
+    Context manager to capture:
+
+        - stdout: replay it, clean it up and make it available via ``obj.out``
+        - stderr: replay it and make it available via ``obj.err``
+
+        init arguments:
+
+        - out - capture stdout:`` True``/``False``, default ``True``
+        - err - capture stdout: ``True``/``False``, default ``True``
+        - replay - whether to replay or not: ``True``/``False``, default ``True``. By default each
+        captured stream gets replayed back on context's exit, so that one can see what the test was
+        doing. If this is a not wanted behavior and the captured data shouldn't be replayed, pass
+        ``replay=False`` to disable this feature.
+
+        Examples::
+
+            # to capture stdout only with auto-replay
+            with CaptureStdout() as cs:
+                print("Secret message")
+            assert "message" in cs.out
+
+            # to capture stderr only with auto-replay
+            import sys
+            with CaptureStderr() as cs:
+                print("Warning: ", file=sys.stderr)
+            assert "Warning" in cs.err
+
+            # to capture both streams with auto-replay
+            with CaptureStd() as cs:
+                print("Secret message")
+                print("Warning: ", file=sys.stderr)
+            assert "message" in cs.out
+            assert "Warning" in cs.err
+
+            # to capture just one of the streams, and not the other, with auto-replay
+            with CaptureStd(err=False) as cs:
+                print("Secret message")
+            assert "message" in cs.out
+            # but best use the stream-specific subclasses
+
+            # to capture without auto-replay
+            with CaptureStd(replay=False) as cs:
+                print("Secret message")
+            assert "message" in cs.out
+
+    """
+
+    def __init__(self, out=True, err=True, replay=True):
+
+        self.replay = replay
+
+        if out:
+            self.out_buf = StringIO()
+            self.out = "error: CaptureStd context is unfinished yet, called too early"
+        else:
+            self.out_buf = None
+            self.out = "not capturing stdout"
+
+        if err:
+            self.err_buf = StringIO()
+            self.err = "error: CaptureStd context is unfinished yet, called too early"
+        else:
+            self.err_buf = None
+            self.err = "not capturing stderr"
+
+    def __enter__(self):
+        if self.out_buf:
+            self.out_old = sys.stdout
+            sys.stdout = self.out_buf
+
+        if self.err_buf:
+            self.err_old = sys.stderr
+            sys.stderr = self.err_buf
+
+        return self
+
+    def __exit__(self, *exc):
+        if self.out_buf:
+            sys.stdout = self.out_old
+            captured = self.out_buf.getvalue()
+            if self.replay:
+                sys.stdout.write(captured)
+            self.out = apply_print_resets(captured)
+
+        if self.err_buf:
+            sys.stderr = self.err_old
+            captured = self.err_buf.getvalue()
+            if self.replay:
+                sys.stderr.write(captured)
+            self.err = captured
+
+    def __repr__(self):
+        msg = ""
+        if self.out_buf:
+            msg += f"stdout: {self.out}\n"
+        if self.err_buf:
+            msg += f"stderr: {self.err}\n"
+        return msg
+
+
+# in tests it's the best to capture only the stream that's wanted, otherwise
+# it's easy to miss things, so unless you need to capture both streams, use the
+# subclasses below (less typing). Or alternatively, configure `CaptureStd` to
+# disable the stream you don't need to test.
+
+
+class CaptureStdout(CaptureStd):
+    """Same as CaptureStd but captures only stdout"""
+
+    def __init__(self, replay=True):
+        super().__init__(err=False, replay=replay)
+
+
+class CaptureStderr(CaptureStd):
+    """Same as CaptureStd but captures only stderr"""
+
+    def __init__(self, replay=True):
+        super().__init__(out=False, replay=replay)
+
+
+class CaptureLogger:
+    """
+    Context manager to capture `logging` streams
+
+    Args:
+
+    - logger: 'logging` logger object
+
+    Results:
+        The captured output is available via `self.out`
+
+    Example::
+
+        >>> from transformers import logging
+        >>> from transformers.testing_utils import CaptureLogger
+
+        >>> msg = "Testing 1, 2, 3"
+        >>> logging.set_verbosity_info()
+        >>> logger = logging.get_logger("transformers.models.bart.tokenization_bart")
+        >>> with CaptureLogger(logger) as cl:
+        ...     logger.info(msg)
+        >>> assert cl.out, msg+"\n"
+    """
+
+    def __init__(self, logger):
+        self.logger = logger
+        self.io = StringIO()
+        self.sh = logging.StreamHandler(self.io)
+        self.out = ""
+
+    def __enter__(self):
+        self.logger.addHandler(self.sh)
+        return self
+
+    def __exit__(self, *exc):
+        self.logger.removeHandler(self.sh)
+        self.out = self.io.getvalue()
+
+    def __repr__(self):
+        return f"captured: {self.out}\n"
+
+
+
+@contextlib.contextmanager
+# adapted from https://stackoverflow.com/a/64789046/9201239
+def ExtendSysPath(path: Union[str, os.PathLike]) -> Iterator[None]:
+    """
+    Temporary add given path to `sys.path`.
+
+    Usage ::
+
+       with ExtendSysPath('/path/to/dir'):
+           mymodule = importlib.import_module('mymodule')
+
+    """
+
+    path = os.fspath(path)
+    try:
+        sys.path.insert(0, path)
+        yield
+    finally:
+        sys.path.remove(path)
+
+
+class TestCasePlus(unittest.TestCase):
+    """
+    This class extends `unittest.TestCase` with additional features.
+
+    Feature 1: A set of fully resolved important file and dir path accessors.
+
+    In tests often we need to know where things are relative to the current test file, and it's not trivial since the
+    test could be invoked from more than one directory or could reside in sub-directories with different depths. This
+    class solves this problem by sorting out all the basic paths and provides easy accessors to them:
+
+    * ``pathlib`` objects (all fully resolved):
+
+       - ``test_file_path`` - the current test file path (=``__file__``)
+       - ``test_file_dir`` - the directory containing the current test file
+       - ``tests_dir`` - the directory of the ``tests`` test suite
+       - ``data_dir`` - the directory of the ``tests/data`` test suite
+       - ``repo_root_dir`` - the directory of the repository
+       - ``src_dir`` - the directory of ``src`` (i.e. where the ``transformers`` sub-dir resides)
+
+    * stringified paths---same as above but these return paths as strings, rather than ``pathlib`` objects:
+
+       - ``test_file_path_str``
+       - ``test_file_dir_str``
+       - ``tests_dir_str``
+       - ``data_dir_str``
+       - ``repo_root_dir_str``
+       - ``src_dir_str``
+
+    Feature 2: Flexible auto-removable temporary dirs which are guaranteed to get removed at the end of test.
+
+    1. Create a unique temporary dir:
+
+    ::
+
+        def test_whatever(self):
+            tmp_dir = self.get_auto_remove_tmp_dir()
+
+    ``tmp_dir`` will contain the path to the created temporary dir. It will be automatically removed at the end of the
+    test.
+
+
+    2. Create a temporary dir of my choice, ensure it's empty before the test starts and don't
+    empty it after the test.
+
+    ::
+
+        def test_whatever(self):
+            tmp_dir = self.get_auto_remove_tmp_dir("./xxx")
+
+    This is useful for debug when you want to monitor a specific directory and want to make sure the previous tests
+    didn't leave any data in there.
+
+    3. You can override the first two options by directly overriding the ``before`` and ``after`` args, leading to the
+       following behavior:
+
+    ``before=True``: the temporary dir will always be cleared at the beginning of the test.
+
+    ``before=False``: if the temporary dir already existed, any existing files will remain there.
+
+    ``after=True``: the temporary dir will always be deleted at the end of the test.
+
+    ``after=False``: the temporary dir will always be left intact at the end of the test.
+
+    Note 1: In order to run the equivalent of ``rm -r`` safely, only subdirs of the project repository checkout are
+    allowed if an explicit ``tmp_dir`` is used, so that by mistake no ``/tmp`` or similar important part of the
+    filesystem will get nuked. i.e. please always pass paths that start with ``./``
+
+    Note 2: Each test can register multiple temporary dirs and they all will get auto-removed, unless requested
+    otherwise.
+
+    Feature 3: Get a copy of the ``os.environ`` object that sets up ``PYTHONPATH`` specific to the current test suite.
+    This is useful for invoking external programs from the test suite - e.g. distributed training.
+
+
+    ::
+        def test_whatever(self):
+            env = self.get_env()
+
+    """
+
+    def setUp(self):
+        # get_auto_remove_tmp_dir feature:
+        self.teardown_tmp_dirs = []
+
+        # figure out the resolved paths for repo_root, tests,  etc.
+        self._test_file_path = inspect.getfile(self.__class__)
+        path = Path(self._test_file_path).resolve()
+        self._test_file_dir = path.parents[0]
+        for up in [1, 2, 3]:
+            tmp_dir = path.parents[up]
+            if (tmp_dir / "megatron").is_dir() and (tmp_dir / "tests").is_dir():
+                break
+        if tmp_dir:
+            self._repo_root_dir = tmp_dir
+        else:
+            raise ValueError(f"can't figure out the root of the repo from {self._test_file_path}")
+        self._tests_dir = self._repo_root_dir / "tests"
+        self._data_dir = self._repo_root_dir / "tests" / "data"
+        self._src_dir = self._repo_root_dir # megatron doesn't use "src/" prefix in the repo
+
+    @property
+    def test_file_path(self):
+        return self._test_file_path
+
+    @property
+    def test_file_path_str(self):
+        return str(self._test_file_path)
+
+    @property
+    def test_file_dir(self):
+        return self._test_file_dir
+
+    @property
+    def test_file_dir_str(self):
+        return str(self._test_file_dir)
+
+    @property
+    def tests_dir(self):
+        return self._tests_dir
+
+    @property
+    def tests_dir_str(self):
+        return str(self._tests_dir)
+
+    @property
+    def data_dir(self):
+        return self._data_dir
+
+    @property
+    def data_dir_str(self):
+        return str(self._data_dir)
+
+    @property
+    def repo_root_dir(self):
+        return self._repo_root_dir
+
+    @property
+    def repo_root_dir_str(self):
+        return str(self._repo_root_dir)
+
+    @property
+    def src_dir(self):
+        return self._src_dir
+
+    @property
+    def src_dir_str(self):
+        return str(self._src_dir)
+
+    def get_env(self):
+        """
+        Return a copy of the ``os.environ`` object that sets up ``PYTHONPATH`` correctly. This is useful
+        for invoking external programs from the test suite - e.g. distributed training.
+
+        It always inserts ``.`` first, then ``./tests`` depending on the test suite type and
+        finally the preset ``PYTHONPATH`` if any (all full resolved paths).
+
+        """
+        env = os.environ.copy()
+        paths = [self.src_dir_str]
+        paths.append(self.tests_dir_str)
+        paths.append(env.get("PYTHONPATH", ""))
+
+        env["PYTHONPATH"] = ":".join(paths)
+        return env
+
+    def get_auto_remove_tmp_dir(self, tmp_dir=None, before=None, after=None):
+        """
+        Args:
+            tmp_dir (:obj:`string`, `optional`):
+                if :obj:`None`:
+
+                   - a unique temporary path will be created
+                   - sets ``before=True`` if ``before`` is :obj:`None`
+                   - sets ``after=True`` if ``after`` is :obj:`None`
+                else:
+
+                   - :obj:`tmp_dir` will be created
+                   - sets ``before=True`` if ``before`` is :obj:`None`
+                   - sets ``after=False`` if ``after`` is :obj:`None`
+            before (:obj:`bool`, `optional`):
+                If :obj:`True` and the :obj:`tmp_dir` already exists, make sure to empty it right away if :obj:`False`
+                and the :obj:`tmp_dir` already exists, any existing files will remain there.
+            after (:obj:`bool`, `optional`):
+                If :obj:`True`, delete the :obj:`tmp_dir` at the end of the test if :obj:`False`, leave the
+                :obj:`tmp_dir` and its contents intact at the end of the test.
+
+        Returns:
+            tmp_dir(:obj:`string`): either the same value as passed via `tmp_dir` or the path to the auto-selected tmp
+            dir
+        """
+        if tmp_dir is not None:
+
+            # defining the most likely desired behavior for when a custom path is provided.
+            # this most likely indicates the debug mode where we want an easily locatable dir that:
+            # 1. gets cleared out before the test (if it already exists)
+            # 2. is left intact after the test
+            if before is None:
+                before = True
+            if after is None:
+                after = False
+
+            # using provided path
+            path = Path(tmp_dir).resolve()
+
+            # to avoid nuking parts of the filesystem, only relative paths are allowed
+            if not tmp_dir.startswith("./"):
+                raise ValueError(
+                    f"`tmp_dir` can only be a relative path, i.e. `./some/path`, but received `{tmp_dir}`"
+                )
+
+            # ensure the dir is empty to start with
+            if before is True and path.exists():
+                shutil.rmtree(tmp_dir, ignore_errors=True)
+
+            path.mkdir(parents=True, exist_ok=True)
+
+        else:
+            # defining the most likely desired behavior for when a unique tmp path is auto generated
+            # (not a debug mode), here we require a unique tmp dir that:
+            # 1. is empty before the test (it will be empty in this situation anyway)
+            # 2. gets fully removed after the test
+            if before is None:
+                before = True
+            if after is None:
+                after = True
+
+            # using unique tmp dir (always empty, regardless of `before`)
+            tmp_dir = tempfile.mkdtemp()
+
+        if after is True:
+            # register for deletion
+            self.teardown_tmp_dirs.append(tmp_dir)
+
+        return tmp_dir
+
+    def tearDown(self):
+
+        # get_auto_remove_tmp_dir feature: remove registered temp dirs
+        for path in self.teardown_tmp_dirs:
+            shutil.rmtree(path, ignore_errors=True)
+        self.teardown_tmp_dirs = []
+
+
+def mockenv(**kwargs):
+    """
+    this is a convenience wrapper, that allows this ::
+
+    @mockenv(RUN_SLOW=True, USE_TF=False)
+    def test_something():
+        run_slow = os.getenv("RUN_SLOW", False)
+        use_tf = os.getenv("USE_TF", False)
+
+    """
+    return mock.patch.dict(os.environ, kwargs)
+
+
+# from https://stackoverflow.com/a/34333710/9201239
+@contextlib.contextmanager
+def mockenv_context(*remove, **update):
+    """
+    Temporarily updates the ``os.environ`` dictionary in-place. Similar to mockenv
+
+    The ``os.environ`` dictionary is updated in-place so that the modification is sure to work in all situations.
+
+    Args:
+      remove: Environment variables to remove.
+      update: Dictionary of environment variables and values to add/update.
+    """
+    env = os.environ
+    update = update or {}
+    remove = remove or []
+
+    # List of environment variables being updated or removed.
+    stomped = (set(update.keys()) | set(remove)) & set(env.keys())
+    # Environment variables and values to restore on exit.
+    update_after = {k: env[k] for k in stomped}
+    # Environment variables and values to remove on exit.
+    remove_after = frozenset(k for k in update if k not in env)
+
+    try:
+        env.update(update)
+        [env.pop(k, None) for k in remove]
+        yield
+    finally:
+        env.update(update_after)
+        [env.pop(k) for k in remove_after]
+
+
+
+# --- distributed testing functions --- #
+
+# adapted from https://stackoverflow.com/a/59041913/9201239
+import asyncio  # noqa
+
+
+class _RunOutput:
+    def __init__(self, returncode, stdout, stderr):
+        self.returncode = returncode
+        self.stdout = stdout
+        self.stderr = stderr
+
+
+async def _read_stream(stream, callback):
+    while True:
+        line = await stream.readline()
+        if line:
+            callback(line)
+        else:
+            break
+
+
+async def _stream_subprocess(cmd, env=None, stdin=None, timeout=None, quiet=False, echo=False) -> _RunOutput:
+    if echo:
+        print("\nRunning: ", " ".join(cmd))
+
+    p = await asyncio.create_subprocess_exec(
+        cmd[0],
+        *cmd[1:],
+        stdin=stdin,
+        stdout=asyncio.subprocess.PIPE,
+        stderr=asyncio.subprocess.PIPE,
+        env=env,
+    )
+
+    # note: there is a warning for a possible deadlock when using `wait` with huge amounts of data in the pipe
+    # https://docs.python.org/3/library/asyncio-subprocess.html#asyncio.asyncio.subprocess.Process.wait
+    #
+    # If it starts hanging, will need to switch to the following code. The problem is that no data
+    # will be seen until it's done and if it hangs for example there will be no debug info.
+    # out, err = await p.communicate()
+    # return _RunOutput(p.returncode, out, err)
+
+    out = []
+    err = []
+
+    def tee(line, sink, pipe, label=""):
+        line = line.decode("utf-8").rstrip()
+        sink.append(line)
+        if not quiet:
+            print(label, line, file=pipe)
+
+    # XXX: the timeout doesn't seem to make any difference here
+    await asyncio.wait(
+        [
+            _read_stream(p.stdout, lambda l: tee(l, out, sys.stdout, label="stdout:")),
+            _read_stream(p.stderr, lambda l: tee(l, err, sys.stderr, label="stderr:")),
+        ],
+        timeout=timeout,
+    )
+    return _RunOutput(await p.wait(), out, err)
+
+
+def execute_subprocess_async(cmd, env=None, stdin=None, timeout=180, quiet=False, echo=True) -> _RunOutput:
+
+    loop = asyncio.get_event_loop()
+    result = loop.run_until_complete(
+        _stream_subprocess(cmd, env=env, stdin=stdin, timeout=timeout, quiet=quiet, echo=echo)
+    )
+
+    cmd_str = " ".join(cmd)
+    if result.returncode > 0:
+        stderr = "\n".join(result.stderr)
+        raise RuntimeError(
+            f"'{cmd_str}' failed with returncode {result.returncode}\n\n"
+            f"The combined stderr from workers follows:\n{stderr}"
+        )
+
+    # check that the subprocess actually did run and produced some output, should the test rely on
+    # the remote side to do the testing
+    if not result.stdout and not result.stderr:
+        raise RuntimeError(f"'{cmd_str}' produced no output.")
+
+    return result
+
+
+# --- Misc utils --- #
+
+def flatten_arguments(args):
+    """
+    Converts dictionary argument to a list.
+
+    Note: we add "IGNORED" at the beginning as this value is ignored by the argparser
+
+    Example: {"arg1": "value1", "arg2": "value2"} -> ["IGNORED", "arg1", "value1", "arg2", "value2"]
+    """
+    return ["IGNORED"] + [item for key_value in args.items() for item in key_value if item != ""]

megatron/text_generation_utils.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Utilities for generating text."""
+
+import copy
+import json
+import os
+import time
+
+import torch
+import torch.nn.functional as F
+
+from megatron import get_args
+from megatron import get_tokenizer
+from megatron import mpu
+from megatron.utils import get_ltor_masks_and_position_ids, unwrap_model
+from megatron.p2p_communication import recv_forward, send_forward
+
+# These are needed to unwrap the model, would be nice to put these in megatron.utils if possible?
+from torch.nn.parallel.distributed import DistributedDataParallel as torchDDP
+from megatron.model import DistributedDataParallel as LocalDDP
+from megatron.model import Float16Module
+
+def get_batch(context_tokens):
+    """Generate batch from context tokens."""
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    # Move to GPU.
+    tokens = context_tokens.view(args.micro_batch_size, -1).contiguous().cuda()
+    # Get the attention mask and position ids.
+    attention_mask, _, position_ids = get_ltor_masks_and_position_ids(
+        tokens,
+        tokenizer.eod,
+        args.reset_position_ids,
+        args.reset_attention_mask,
+        args.eod_mask_loss,
+        prefix_indices=None,
+        loss_on_targets_only=args.loss_on_targets_only
+    )
+
+    return tokens, attention_mask, position_ids
+
+
+def top_k_logits(logits, top_k=0, top_p=0.0, filter_value=-float('Inf')):
+    """ This function has been mostly taken from huggingface conversational
+     ai code at
+         https://medium.com/huggingface/how-to-build-a-state-of-the-art-
+              conversational-ai-with-transfer-learning-2d818ac26313 """
+
+    if top_k > 0:
+        # Remove all tokens with a probability less than the
+        # last token of the top-k
+        indices_to_remove = logits < torch.topk(logits, top_k)[0][..., -1, None]
+        logits[indices_to_remove] = filter_value
+
+    if top_p > 0.0:
+        # Cconvert to 1D
+        sorted_logits, sorted_indices = torch.sort(
+            logits, descending=True, dim=-1)
+        cumulative_probs = torch.cumsum(F.softmax(sorted_logits, dim=-1),
+                                        dim=-1)
+
+        # Remove tokens with cumulative probability above the threshold
+        sorted_indices_to_remove = cumulative_probs > top_p
+        # Shift the indices to the right to keep also the first token
+        # above the threshold
+        sorted_indices_to_remove[..., 1:] \
+            = sorted_indices_to_remove[..., :-1].clone()
+        sorted_indices_to_remove[..., 0] = 0
+        for i in range(sorted_indices.size(0)):
+            indices_to_remove = sorted_indices[i][sorted_indices_to_remove[i]]
+            logits[i][indices_to_remove] = filter_value
+
+    return logits
+
+
+def generate_samples_input_from_file(model):
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    # Read the sample file and open the output file.
+    assert args.sample_input_file is not None, \
+        'sample input file is not provided.'
+    if mpu.is_pipeline_first_stage() and mpu.get_tensor_model_parallel_rank() == 0:
+        fname = open(args.sample_input_file, "r")
+        all_raw_text = fname.readlines()
+        input_count = len(all_raw_text)
+        input_pos = 0
+        if args.sample_output_file is None:
+            sample_output_file = args.sample_input_file + ".out"
+            print('`sample-output-file` not specified, setting '
+                  'it to {}'.format(sample_output_file))
+        else:
+            sample_output_file = args.sample_output_file
+        fname_out = open(sample_output_file, "w+")
+
+    context_count = 0
+    model.eval()
+    with torch.no_grad():
+        while True:
+            terminate_runs = 0
+            raw_text_len = 0
+
+            if mpu.is_pipeline_first_stage() \
+               and mpu.get_tensor_model_parallel_rank() == 0:
+                raw_text = all_raw_text[input_pos]
+                input_pos += 1
+                if input_pos == input_count:
+                    raw_text = "stop"
+                raw_text_len = len(raw_text)
+
+                if "stop" in raw_text:
+                    terminate_runs = 1
+                else:
+                    context_tokens = tokenizer.tokenize(raw_text)
+                    context_length = len(context_tokens)
+
+                    if context_length >= (args.seq_length // 2):
+                        print("\nContext length", context_length,
+                              "\nPlease give smaller context (half of the "
+                              "sequence length)!", flush=True)
+                        continue
+            else:
+                context_tokens = tokenizer.tokenize("EMPTY TEXT")
+                context_length = 0
+
+            input_info = [terminate_runs, raw_text_len, context_length]
+            input_info_tensor = torch.cuda.LongTensor(input_info)
+            torch.distributed.all_reduce(input_info_tensor,
+                                         group=mpu.get_model_parallel_group())
+            terminate_runs = input_info_tensor[0].item()
+            raw_text_len = input_info_tensor[1].item()
+            context_length = input_info_tensor[2].item()
+
+            if terminate_runs == 1:
+                return
+
+            # For pipeline parallel we send context tokens to other stages
+            # so they get the lengths correct
+            if mpu.get_tensor_model_parallel_rank() == 0 \
+               and args.pipeline_model_parallel_size > 1:
+                if mpu.is_pipeline_first_stage():
+                    src = mpu.get_pipeline_model_parallel_first_rank()
+                    group = mpu.get_pipeline_model_parallel_group()
+                    context_tokens_tensor = torch.cuda.LongTensor(context_tokens)
+                    torch.distributed.broadcast(context_tokens_tensor, src, group)
+                else:
+                    src = mpu.get_pipeline_model_parallel_first_rank()
+                    group = mpu.get_pipeline_model_parallel_group()
+                    context_tokens_tensor = torch.empty(context_length,
+                                                        dtype=torch.int64,
+                                                        device=torch.device("cuda"))
+                    torch.distributed.broadcast(context_tokens_tensor, src, group)
+                    context_tokens = context_tokens_tensor.cpu().numpy().tolist()
+
+            token_stream = get_token_stream(model, [context_tokens])
+            for _, decode_tokens in enumerate(token_stream):
+                pass
+
+            if mpu.get_tensor_model_parallel_rank() == 0:
+                if mpu.is_pipeline_first_stage():
+                    os.system('clear')
+                    print("\nContext:", raw_text, flush=True)
+
+                    fname_out.write("\nContext:")
+                    fname_out.write(raw_text)
+
+                    decode_tokens, _ = decode_tokens
+                    decode_tokens = decode_tokens[0].cpu().numpy().tolist()
+                    trim_decode_tokens = tokenizer.detokenize(
+                        decode_tokens)[raw_text_len:]
+                    print("\nMegatron-LM:", trim_decode_tokens, flush=True)
+
+                    fname_out.write("\n\nMegatron-LM:")
+                    fname_out.write(trim_decode_tokens)
+                    fname_out.write("\n")
+
+            raw_text = None
+            context_count += 1
+
+# We added this function to support the tasks evaluation such as squad
+# and drop in the https://github.com/EleutherAI/lm-evaluation-harness 
+# codebase. The lm-evaluation-harness code can now call this function
+# similar to their current generate function call used for gpt style models.
+def generate_samples_eval(model, context, max_gen_length, eos_token_id):
+    # Generate samples for lm evaluation
+    # NEED TO THINK ABOUT eos token
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    raw_text_len = len(context)
+    model.eval()
+
+    context_tokens = tokenizer.tokenize(context)
+    args.out_seq_length = max_gen_length + len(context_tokens)
+    args.eos_id = eos_token_id
+
+    with torch.no_grad():
+        token_stream = get_token_stream(model, [context_tokens])
+        for counter, decode_tokens in enumerate(token_stream):
+            if counter == args.out_seq_length:
+                break
+
+    decode_tokens, _ = decode_tokens
+    decode_tokens = decode_tokens[0].cpu().numpy().tolist()
+    trim_decode_tokens = tokenizer.detokenize(
+        decode_tokens)[raw_text_len:]
+ 
+    return trim_decode_tokens
+
+
+def generate_samples_interactive(model, print_frequency=24):
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    context_count = 0
+    model.eval()
+    with torch.no_grad():
+        while True:
+            terminate_runs = 0
+            raw_text_len = 0
+
+            if mpu.is_pipeline_first_stage() \
+               and mpu.get_tensor_model_parallel_rank() == 0:
+                os.system('clear')
+                raw_text = input("\nContext prompt (stop to exit) >>> ")
+                while not raw_text:
+                    print('Prompt should not be empty!')
+                    raw_text = input("\nContext prompt (stop to exit) >>> ")
+                raw_text_len = len(raw_text)
+
+                if "stop" in raw_text:
+                    terminate_runs = 1
+                else:
+                    context_tokens = tokenizer.tokenize(raw_text)
+                    context_length = len(context_tokens)
+
+                    if context_length >= (args.seq_length // 2):
+                        print("\nContext length", context_length,
+                              "\nPlease give smaller context (half of the "
+                              "sequence length)!", flush=True)
+                        continue
+            else:
+                context_tokens = tokenizer.tokenize("EMPTY TEXT")
+                context_length = 0
+
+            input_info = [terminate_runs, raw_text_len, context_length]
+            input_info_tensor = torch.cuda.LongTensor(input_info)
+            torch.distributed.all_reduce(input_info_tensor,
+                                         group=mpu.get_model_parallel_group())
+            terminate_runs = input_info_tensor[0].item()
+            raw_text_len = input_info_tensor[1].item()
+            context_length = input_info_tensor[2].item()
+
+            if terminate_runs == 1:
+                return
+
+            # For pipeline parallel we send context tokens to other stages
+            # so they get the lengths correct
+            if mpu.get_tensor_model_parallel_rank() == 0 \
+               and args.pipeline_model_parallel_size > 1:
+                if mpu.is_pipeline_first_stage():
+                    src = mpu.get_pipeline_model_parallel_first_rank()
+                    group = mpu.get_pipeline_model_parallel_group()
+                    context_tokens_tensor = torch.cuda.LongTensor(context_tokens)
+                    torch.distributed.broadcast(context_tokens_tensor, src, group)
+                else:
+                    src = mpu.get_pipeline_model_parallel_first_rank()
+                    group = mpu.get_pipeline_model_parallel_group()
+                    context_tokens_tensor = torch.empty(context_length,
+                                                        dtype=torch.int64,
+                                                        device=torch.device("cuda"))
+                    torch.distributed.broadcast(context_tokens_tensor, src, group)
+                    context_tokens = context_tokens_tensor.cpu().numpy().tolist()
+
+            token_stream = get_token_stream(model, [context_tokens])
+
+            for counter, decode_tokens in enumerate(token_stream):
+                if counter % print_frequency != 0 \
+                   or mpu.get_tensor_model_parallel_rank() != 0 \
+                   or not mpu.is_pipeline_first_stage():
+                    continue
+
+                os.system('clear')
+                print("\nContext:", raw_text, flush=True)
+
+                decode_tokens, _ = decode_tokens
+                decode_tokens = decode_tokens[0].cpu().numpy().tolist()
+                trim_decode_tokens = tokenizer.detokenize(
+                    decode_tokens)[raw_text_len:]
+                print("\nMegatron-LM:", trim_decode_tokens, flush=True)
+
+            if mpu.is_pipeline_first_stage() \
+               and mpu.get_tensor_model_parallel_rank() == 0:
+                os.system('clear')
+                print("\nContext:", raw_text, flush=True)
+
+                if not isinstance(decode_tokens, list):
+                    decode_tokens, _ = decode_tokens
+                    decode_tokens = decode_tokens[0].cpu().numpy().tolist()
+                trim_decode_tokens = tokenizer.detokenize(
+                    decode_tokens)[raw_text_len:]
+                print("\nMegatron-LM:", trim_decode_tokens, flush=True)
+
+                input("\nPress Enter to continue >>>")
+
+            raw_text = None
+            context_count += 1
+
+
+
+def generate_samples_unconditional(model):
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    num_samples = args.num_samples
+    context_tokens = [[tokenizer.eod]
+                      for _ in range(args.micro_batch_size)]
+    ctr = 0
+    while True:
+        start_time = time.time()
+        for token_stream in get_token_stream(model,
+                                             copy.deepcopy(context_tokens)):
+            pass
+        if mpu.is_pipeline_last_stage() and \
+           mpu.get_tensor_model_parallel_rank() == 0:
+            if ctr % args.log_interval == 0:
+                print('Avg s/batch:',
+                      (time.time() - start_time) / min(args.log_interval, ctr + 1))
+                start_time = time.time()
+            length = len(token_stream)
+            token_batch = token_stream[0].cpu().numpy().tolist()
+            length_batch = token_stream[1].cpu().numpy().tolist()
+            assert len(length_batch) == args.micro_batch_size
+            for tokens, length in zip(token_batch, length_batch):
+                tokens = tokens[1:length - 1]
+                text = tokenizer.detokenize(tokens)
+                is_finished = length < args.seq_length - 1
+                datum = {'text': text, 'length': length - 1, 'finished': is_finished}
+                yield datum
+                ctr += 1
+                if ctr >= num_samples:
+                    break
+        else:
+            for _ in range(args.micro_batch_size):
+                yield None
+                ctr += 1
+                if ctr >= num_samples:
+                    break
+        if ctr >= num_samples:
+            break
+
+
+def generate_and_write_samples_unconditional(model):
+
+    args = get_args()
+    assert args.genfile is not None
+    with open(args.genfile, 'w') as f:
+        for datum in generate_samples_unconditional(model):
+            if mpu.is_pipeline_last_stage() and \
+               mpu.get_tensor_model_parallel_rank() == 0:
+                f.write(json.dumps(datum) + '\n')
+
+
+def pad_batch(batch, pad_id, args):
+
+    context_lengths = []
+    for tokens in batch:
+        context_length = len(tokens)
+        if context_length < args.seq_length:
+            tokens.extend([pad_id] * (args.seq_length - context_length))
+        context_lengths.append(context_length)
+    return batch, context_lengths
+
+
+def get_token_stream(model, context_tokens):
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    context_tokens, context_lengths = pad_batch(context_tokens,
+                                                tokenizer.eod, args)
+
+    context_tokens_tensor = torch.cuda.LongTensor(context_tokens)
+    context_length_tensor = torch.cuda.LongTensor(context_lengths)
+
+    torch.distributed.broadcast(context_length_tensor,
+                                mpu.get_tensor_model_parallel_src_rank(),
+                                group=mpu.get_tensor_model_parallel_group())
+    torch.distributed.broadcast(context_tokens_tensor,
+                                mpu.get_tensor_model_parallel_src_rank(),
+                                group=mpu.get_tensor_model_parallel_group())
+
+    context_length = context_length_tensor.min().item()
+    tokens, attention_mask, position_ids = get_batch(context_tokens_tensor)
+
+    batch_token_iterator = sample_sequence_batch(model, context_tokens_tensor,
+                                                 context_length_tensor,
+                                                 attention_mask, position_ids)
+    for tokens, lengths in batch_token_iterator:
+        context_length += 1
+        if tokens is not None:
+            yield tokens[:, :context_length], lengths
+        else:
+            yield None, None
+
+
+def switch(val1, val2, boolean):
+
+    boolean = boolean.type_as(val1)
+    return (1 - boolean) * val1 + boolean * val2
+
+
+def forward_step(model, tokens, position_ids, attention_mask, tokentype_ids,
+                 layer_past=None, get_key_value=None,
+                 forward_method_parallel_output=None):
+
+    # Hidden size changes when not using recompute, need to tell p2p_communicate
+    # functions the correct size
+    args = get_args()
+    orig_seq_length = args.seq_length
+    args.seq_length = tokens.shape[1]
+
+    input_tensor = recv_forward()
+
+    # Forward pass through the model.
+    unwrapped_model = unwrap_model(
+        model, (torchDDP, LocalDDP, Float16Module))
+    unwrapped_model.set_input_tensor(input_tensor)
+    output_tensor = model(tokens, position_ids, attention_mask,
+                          tokentype_ids=tokentype_ids,
+                          layer_past=layer_past,
+                          get_key_value=get_key_value,
+                          forward_method_parallel_output=forward_method_parallel_output)
+
+    if get_key_value:
+        output_tensor, layer_past = output_tensor
+
+    send_forward(output_tensor)
+
+    args.seq_length = orig_seq_length
+    if get_key_value:
+        return output_tensor, layer_past
+    return output_tensor
+
+
+def sample_sequence_batch(model, context_tokens, context_lengths,
+                          attention_mask, position_ids,
+                          maxlen=None, type_ids=None):
+
+    args = get_args()
+    tokenizer = get_tokenizer()
+
+    model.eval()
+    with torch.no_grad():
+        context_length = context_lengths.min().item()
+
+        # added eos_id to support the function generate_samples_eval that passes
+        # eos_id as an argument and needs termination when that id id found.
+        if hasattr(args, 'eos_id'):
+            eos_id = args.eos_id
+        else:
+            eos_id = tokenizer.eod
+
+        counter = 0
+        org_context_length = context_length
+
+        layer_past = None
+        batch_size = context_tokens.size(0)
+        is_done = torch.zeros([batch_size]).byte().cuda()
+        tokens = context_tokens
+        if maxlen is None:
+            maxlen = args.seq_length - 1
+            if maxlen > (org_context_length + args.out_seq_length):
+                maxlen = org_context_length + args.out_seq_length
+
+        lengths = torch.ones([batch_size]).long().cuda() * maxlen
+
+        while context_length <= (maxlen):
+            if args.recompute:
+                output = forward_step(model, tokens,
+                                      position_ids,
+                                      attention_mask,
+                                      tokentype_ids=type_ids,
+                                      forward_method_parallel_output=False)
+                if mpu.is_pipeline_last_stage():
+                    assert output is not None
+                    logits = output[:, context_length - 1, :]
+            else:
+                types2use = None
+                if counter == 0:
+                    tokens2use = tokens[:, :context_length]
+                    positions2use = position_ids[:, :context_length]
+                    if type_ids is not None:
+                        types2use = type_ids[:, :context_length]
+                else:
+                    tokens2use = tokens[:, context_length - 1].view(
+                        batch_size, -1)
+                    positions2use = position_ids[:, context_length - 1].view(
+                        batch_size, -1)
+                    if type_ids is not None:
+                        types2use = type_ids[:, context_length - 1].view(
+                            batch_size, -1)
+                output, layer_past = forward_step(model, tokens2use,
+                                                  positions2use,
+                                                  attention_mask,
+                                                  layer_past=layer_past,
+                                                  get_key_value=True,
+                                                  tokentype_ids=types2use,
+                                                  forward_method_parallel_output=False)
+                if mpu.is_pipeline_last_stage():
+                    assert output is not None
+                    logits = output[:, -1].view(batch_size, -1).contiguous()
+
+            if mpu.is_pipeline_last_stage():
+                if args.greedy:
+                    prev = torch.argmax(logits, dim=-1).view(-1)
+                else:
+                    logits = logits.float()
+                    logits /= args.temperature
+                    logits = top_k_logits(logits, top_k=args.top_k,
+                                          top_p=args.top_p)
+                    log_probs = F.softmax(logits, dim=-1)
+                    prev = torch.multinomial(log_probs, num_samples=1).view(-1)
+
+                started = context_lengths <= context_length
+
+                new_tokens = switch(
+                    tokens[:, context_length].view(-1), prev, started)
+                tokens[:, context_length] = new_tokens
+                src = mpu.get_pipeline_model_parallel_last_rank()
+                group = mpu.get_embedding_group()
+                torch.distributed.broadcast(new_tokens, src, group)
+
+                done_token = (prev == eos_id).byte() & started.byte()
+                just_finished = (done_token & ~is_done).bool()
+                lengths[just_finished.view(-1)] = context_length
+                is_done = is_done | done_token
+
+                done = torch.all(is_done)
+                src = mpu.get_pipeline_model_parallel_last_rank()
+                group = mpu.get_pipeline_model_parallel_group()
+                torch.distributed.broadcast(done, src, group)
+                yield tokens, lengths
+
+            else:
+                if mpu.is_pipeline_first_stage():
+                    src = mpu.get_pipeline_model_parallel_last_rank()
+                    group = mpu.get_embedding_group()
+                    new_tokens = torch.empty_like(tokens[:, context_length])
+                    torch.distributed.broadcast(new_tokens, src, group)
+                    tokens[:, context_length] = new_tokens
+                    yield tokens, None
+                else:
+                    yield None, None
+
+                done = torch.cuda.ByteTensor([0])
+                src = mpu.get_pipeline_model_parallel_last_rank()
+                group = mpu.get_pipeline_model_parallel_group()
+                torch.distributed.broadcast(done, src, group)
+
+            context_length += 1
+            counter += 1
+            if done:
+                break

megatron/tokenizer/__init__.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+from .tokenizer import build_tokenizer

megatron/tokenizer/bert_tokenization.py

+# coding=utf-8
+# Copyright 2018 The Google AI Language Team Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Tokenization classes."""
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import collections
+import re
+import unicodedata
+import six
+
+
+def validate_case_matches_checkpoint(do_lower_case, init_checkpoint):
+    """Checks whether the casing config is consistent with the checkpoint name."""
+
+    # The casing has to be passed in by the user and there is no explicit check
+    # as to whether it matches the checkpoint. The casing information probably
+    # should have been stored in the bert_config.json file, but it's not, so
+    # we have to heuristically detect it to validate.
+
+    if not init_checkpoint:
+        return
+
+    m = re.match("^.*?([A-Za-z0-9_-]+)/bert_model.ckpt", init_checkpoint)
+    if m is None:
+        return
+
+    model_name = m.group(1)
+
+    lower_models = [
+        "uncased_L-24_H-1024_A-16", "uncased_L-12_H-768_A-12",
+        "multilingual_L-12_H-768_A-12", "chinese_L-12_H-768_A-12"
+    ]
+
+    cased_models = [
+        "cased_L-12_H-768_A-12", "cased_L-24_H-1024_A-16",
+        "multi_cased_L-12_H-768_A-12"
+    ]
+
+    is_bad_config = False
+    if model_name in lower_models and not do_lower_case:
+        is_bad_config = True
+        actual_flag = "False"
+        case_name = "lowercased"
+        opposite_flag = "True"
+
+    if model_name in cased_models and do_lower_case:
+        is_bad_config = True
+        actual_flag = "True"
+        case_name = "cased"
+        opposite_flag = "False"
+
+    if is_bad_config:
+        raise ValueError(
+            "You passed in `--do_lower_case=%s` with `--init_checkpoint=%s`. "
+            "However, `%s` seems to be a %s model, so you "
+            "should pass in `--do_lower_case=%s` so that the fine-tuning matches "
+            "how the model was pre-training. If this error is wrong, please "
+            "just comment out this check." % (actual_flag, init_checkpoint,
+                                              model_name, case_name, opposite_flag))
+
+
+def convert_to_unicode(text):
+    """Converts `text` to Unicode (if it's not already), assuming utf-8 input."""
+    if six.PY3:
+        if isinstance(text, str):
+            return text
+        elif isinstance(text, bytes):
+            return text.decode("utf-8", "ignore")
+        else:
+            raise ValueError("Unsupported string type: %s" % (type(text)))
+    elif six.PY2:
+        if isinstance(text, str):
+            return text.decode("utf-8", "ignore")
+        elif isinstance(text, unicode):
+            return text
+        else:
+            raise ValueError("Unsupported string type: %s" % (type(text)))
+    else:
+        raise ValueError("Not running on Python2 or Python 3?")
+
+
+def printable_text(text):
+    """Returns text encoded in a way suitable for print or `tf.logging`."""
+
+    # These functions want `str` for both Python2 and Python3, but in one case
+    # it's a Unicode string and in the other it's a byte string.
+    if six.PY3:
+        if isinstance(text, str):
+            return text
+        elif isinstance(text, bytes):
+            return text.decode("utf-8", "ignore")
+        else:
+            raise ValueError("Unsupported string type: %s" % (type(text)))
+    elif six.PY2:
+        if isinstance(text, str):
+            return text
+        elif isinstance(text, unicode):
+            return text.encode("utf-8")
+        else:
+            raise ValueError("Unsupported string type: %s" % (type(text)))
+    else:
+        raise ValueError("Not running on Python2 or Python 3?")
+
+
+def load_vocab(vocab_file):
+    """Loads a vocabulary file into a dictionary."""
+    vocab = collections.OrderedDict()
+    index = 0
+    with open(vocab_file, "r") as reader:
+        while True:
+            token = convert_to_unicode(reader.readline())
+            if not token:
+                break
+            token = token.strip()
+            vocab[token] = index
+            index += 1
+    return vocab
+
+
+def convert_by_vocab(vocab, items):
+    """Converts a sequence of [tokens|ids] using the vocab."""
+    output = []
+    for item in items:
+        output.append(vocab[item])
+    return output
+
+
+def convert_tokens_to_ids(vocab, tokens):
+    return convert_by_vocab(vocab, tokens)
+
+
+def convert_ids_to_tokens(inv_vocab, ids):
+    return convert_by_vocab(inv_vocab, ids)
+
+
+def whitespace_tokenize(text):
+    """Runs basic whitespace cleaning and splitting on a piece of text."""
+    text = text.strip()
+    if not text:
+        return []
+    tokens = text.split()
+    return tokens
+
+
+class FullTokenizer(object):
+    """Runs end-to-end tokenziation."""
+
+    def __init__(self, vocab_file, do_lower_case=True):
+        self.vocab = load_vocab(vocab_file)
+        self.inv_vocab = {v: k for k, v in self.vocab.items()}
+        self.basic_tokenizer = BasicTokenizer(do_lower_case=do_lower_case)
+        self.wordpiece_tokenizer = WordpieceTokenizer(vocab=self.vocab)
+
+    def tokenize(self, text):
+        split_tokens = []
+        for token in self.basic_tokenizer.tokenize(text):
+            for sub_token in self.wordpiece_tokenizer.tokenize(token):
+                split_tokens.append(sub_token)
+
+        return split_tokens
+
+    def convert_tokens_to_ids(self, tokens):
+        return convert_by_vocab(self.vocab, tokens)
+
+    def convert_ids_to_tokens(self, ids):
+        return convert_by_vocab(self.inv_vocab, ids)
+
+    def vocab_size(self):
+        return len(self.vocab)
+
+
+class BasicTokenizer(object):
+    """Runs basic tokenization (punctuation splitting, lower casing, etc.)."""
+
+    def __init__(self, do_lower_case=True):
+        """Constructs a BasicTokenizer.
+
+        Args:
+          do_lower_case: Whether to lower case the input.
+        """
+        self.do_lower_case = do_lower_case
+
+    def tokenize(self, text):
+        """Tokenizes a piece of text."""
+        text = convert_to_unicode(text)
+        text = self._clean_text(text)
+
+        # This was added on November 1st, 2018 for the multilingual and Chinese
+        # models. This is also applied to the English models now, but it doesn't
+        # matter since the English models were not trained on any Chinese data
+        # and generally don't have any Chinese data in them (there are Chinese
+        # characters in the vocabulary because Wikipedia does have some Chinese
+        # words in the English Wikipedia.).
+        text = self._tokenize_chinese_chars(text)
+
+        orig_tokens = whitespace_tokenize(text)
+        split_tokens = []
+        for token in orig_tokens:
+            if self.do_lower_case:
+                token = token.lower()
+                token = self._run_strip_accents(token)
+            split_tokens.extend(self._run_split_on_punc(token))
+
+        output_tokens = whitespace_tokenize(" ".join(split_tokens))
+        return output_tokens
+
+    def _run_strip_accents(self, text):
+        """Strips accents from a piece of text."""
+        text = unicodedata.normalize("NFD", text)
+        output = []
+        for char in text:
+            cat = unicodedata.category(char)
+            if cat == "Mn":
+                continue
+            output.append(char)
+        return "".join(output)
+
+    def _run_split_on_punc(self, text):
+        """Splits punctuation on a piece of text."""
+        chars = list(text)
+        i = 0
+        start_new_word = True
+        output = []
+        while i < len(chars):
+            char = chars[i]
+            if _is_punctuation(char):
+                output.append([char])
+                start_new_word = True
+            else:
+                if start_new_word:
+                    output.append([])
+                start_new_word = False
+                output[-1].append(char)
+            i += 1
+
+        return ["".join(x) for x in output]
+
+    def _tokenize_chinese_chars(self, text):
+        """Adds whitespace around any CJK character."""
+        output = []
+        for char in text:
+            cp = ord(char)
+            if self._is_chinese_char(cp):
+                output.append(" ")
+                output.append(char)
+                output.append(" ")
+            else:
+                output.append(char)
+        return "".join(output)
+
+    def _is_chinese_char(self, cp):
+        """Checks whether CP is the codepoint of a CJK character."""
+        # This defines a "chinese character" as anything in the CJK Unicode block:
+        #   https://en.wikipedia.org/wiki/CJK_Unified_Ideographs_(Unicode_block)
+        #
+        # Note that the CJK Unicode block is NOT all Japanese and Korean characters,
+        # despite its name. The modern Korean Hangul alphabet is a different block,
+        # as is Japanese Hiragana and Katakana. Those alphabets are used to write
+        # space-separated words, so they are not treated specially and handled
+        # like the all of the other languages.
+        if ((cp >= 0x4E00 and cp <= 0x9FFF) or  #
+            (cp >= 0x3400 and cp <= 0x4DBF) or  #
+            (cp >= 0x20000 and cp <= 0x2A6DF) or  #
+            (cp >= 0x2A700 and cp <= 0x2B73F) or  #
+            (cp >= 0x2B740 and cp <= 0x2B81F) or  #
+            (cp >= 0x2B820 and cp <= 0x2CEAF) or
+            (cp >= 0xF900 and cp <= 0xFAFF) or  #
+                (cp >= 0x2F800 and cp <= 0x2FA1F)):  #
+            return True
+
+        return False
+
+    def _clean_text(self, text):
+        """Performs invalid character removal and whitespace cleanup on text."""
+        output = []
+        for char in text:
+            cp = ord(char)
+            if cp == 0 or cp == 0xfffd or _is_control(char):
+                continue
+            if _is_whitespace(char):
+                output.append(" ")
+            else:
+                output.append(char)
+        return "".join(output)
+
+
+class WordpieceTokenizer(object):
+    """Runs WordPiece tokenziation."""
+
+    def __init__(self, vocab, unk_token="[UNK]", max_input_chars_per_word=200):
+        self.vocab = vocab
+        self.unk_token = unk_token
+        self.max_input_chars_per_word = max_input_chars_per_word
+
+    def tokenize(self, text):
+        """Tokenizes a piece of text into its word pieces.
+
+        This uses a greedy longest-match-first algorithm to perform tokenization
+        using the given vocabulary.
+
+        For example:
+          input = "unaffable"
+          output = ["un", "##aff", "##able"]
+
+        Args:
+          text: A single token or whitespace separated tokens. This should have
+            already been passed through `BasicTokenizer.
+
+        Returns:
+          A list of wordpiece tokens.
+        """
+
+        text = convert_to_unicode(text)
+
+        output_tokens = []
+        for token in whitespace_tokenize(text):
+            chars = list(token)
+            if len(chars) > self.max_input_chars_per_word:
+                output_tokens.append(self.unk_token)
+                continue
+
+            is_bad = False
+            start = 0
+            sub_tokens = []
+            while start < len(chars):
+                end = len(chars)
+                cur_substr = None
+                while start < end:
+                    substr = "".join(chars[start:end])
+                    if start > 0:
+                        substr = "##" + substr
+                    if substr in self.vocab:
+                        cur_substr = substr
+                        break
+                    end -= 1
+                if cur_substr is None:
+                    is_bad = True
+                    break
+                sub_tokens.append(cur_substr)
+                start = end
+
+            if is_bad:
+                output_tokens.append(self.unk_token)
+            else:
+                output_tokens.extend(sub_tokens)
+        return output_tokens
+
+
+def _is_whitespace(char):
+    """Checks whether `chars` is a whitespace character."""
+    # \t, \n, and \r are technically contorl characters but we treat them
+    # as whitespace since they are generally considered as such.
+    if char == " " or char == "\t" or char == "\n" or char == "\r":
+        return True
+    cat = unicodedata.category(char)
+    if cat == "Zs":
+        return True
+    return False
+
+
+def _is_control(char):
+    """Checks whether `chars` is a control character."""
+    # These are technically control characters but we count them as whitespace
+    # characters.
+    if char == "\t" or char == "\n" or char == "\r":
+        return False
+    cat = unicodedata.category(char)
+    if cat in ("Cc", "Cf"):
+        return True
+    return False
+
+
+def _is_punctuation(char):
+    """Checks whether `chars` is a punctuation character."""
+    cp = ord(char)
+    # We treat all non-letter/number ASCII as punctuation.
+    # Characters such as "^", "$", and "`" are not in the Unicode
+    # Punctuation class but we treat them as punctuation anyways, for
+    # consistency.
+    if ((cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or
+            (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126)):
+        return True
+    cat = unicodedata.category(char)
+    if cat.startswith("P"):
+        return True
+    return False

megatron/tokenizer/gpt2_tokenization.py

+# coding=utf-8
+# Copyright 2018 The Open AI Team Authors and The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Tokenization classes for OpenAI GPT."""
+
+from __future__ import (absolute_import, division, print_function,
+                        unicode_literals)
+
+import sys
+import json
+import os
+import regex as re
+from io import open
+
+from megatron import logging
+
+try:
+    from functools import lru_cache
+except ImportError:
+    # Just a dummy decorator to get the checks to run on python2
+    # because honestly I don't want to support a byte-level unicode BPE
+    # tokenizer on python 2 right now.
+    def lru_cache():
+        return lambda func: func
+
+
+logger = logging.get_logger(__name__)
+
+PRETRAINED_VOCAB_ARCHIVE_MAP = {
+    'gpt2': "https://s3.amazonaws.com/models.huggingface.co/bert/gpt2-vocab.json",
+}
+PRETRAINED_MERGES_ARCHIVE_MAP = {
+    'gpt2': "https://s3.amazonaws.com/models.huggingface.co/bert/gpt2-merges.txt",
+}
+PRETRAINED_VOCAB_POSITIONAL_EMBEDDINGS_SIZE_MAP = {
+    'gpt2': 1024,
+}
+VOCAB_NAME = 'vocab.json'
+MERGES_NAME = 'merges.txt'
+SPECIAL_TOKENS_NAME = 'special_tokens.txt'
+
+
+@lru_cache()
+def bytes_to_unicode():
+    """
+    Returns list of utf-8 byte and a corresponding list of unicode strings.
+    The reversible bpe codes work on unicode strings.
+    This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
+    When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
+    This is a signficant percentage of your normal, say, 32K bpe vocab.
+    To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
+    And avoids mapping to whitespace/control characters the bpe code barfs on.
+    """
+    _chr = unichr if sys.version_info[0] == 2 else chr
+    bs = list(range(ord("!"), ord("~") + 1)) + list(range(ord("¡"), ord("¬") + 1)) + \
+        list(range(ord("®"), ord("ÿ") + 1))
+    cs = bs[:]
+    n = 0
+    for b in range(2**8):
+        if b not in bs:
+            bs.append(b)
+            cs.append(2**8 + n)
+            n += 1
+    cs = [_chr(n) for n in cs]
+    return dict(zip(bs, cs))
+
+
+def get_pairs(word):
+    """Return set of symbol pairs in a word.
+    Word is represented as tuple of symbols (symbols being variable-length strings).
+    """
+    pairs = set()
+    prev_char = word[0]
+    for char in word[1:]:
+        pairs.add((prev_char, char))
+        prev_char = char
+    return pairs
+
+
+class GPT2Tokenizer(object):
+    """
+    GPT-2 BPE tokenizer. Peculiarities:
+        - Byte-level BPE
+    """
+    @classmethod
+    def from_pretrained(cls, pretrained_model_name_or_path, cache_dir=None, *inputs, **kwargs):
+        """
+        Instantiate a PreTrainedBertModel from a pre-trained model file.
+        Download and cache the pre-trained model file if needed.
+        """
+        if pretrained_model_name_or_path in PRETRAINED_VOCAB_ARCHIVE_MAP:
+            vocab_file = PRETRAINED_VOCAB_ARCHIVE_MAP[pretrained_model_name_or_path]
+            merges_file = PRETRAINED_MERGES_ARCHIVE_MAP[pretrained_model_name_or_path]
+            special_tokens_file = None
+        else:
+            vocab_file = os.path.join(pretrained_model_name_or_path, VOCAB_NAME)
+            merges_file = os.path.join(pretrained_model_name_or_path, MERGES_NAME)
+            special_tokens_file = os.path.join(pretrained_model_name_or_path, SPECIAL_TOKENS_NAME)
+            if not os.path.exists(special_tokens_file):
+                special_tokens_file = None
+            else:
+                logger.info("loading special tokens file {}".format(special_tokens_file))
+        # redirect to the cache, if necessary
+        try:
+            from .file_utils import cached_path
+            resolved_vocab_file = cached_path(vocab_file, cache_dir=cache_dir)
+            resolved_merges_file = cached_path(merges_file, cache_dir=cache_dir)
+        except EnvironmentError:
+            logger.error(
+                "Model name '{}' was not found in model name list ({}). "
+                "We assumed '{}' was a path or url but couldn't find files {} and {} "
+                "at this path or url.".format(
+                    pretrained_model_name_or_path,
+                    ', '.join(PRETRAINED_VOCAB_ARCHIVE_MAP.keys()),
+                    pretrained_model_name_or_path,
+                    vocab_file, merges_file))
+            return None
+        if resolved_vocab_file == vocab_file and resolved_merges_file == merges_file:
+            logger.info("loading vocabulary file {}".format(vocab_file))
+            logger.info("loading merges file {}".format(merges_file))
+        else:
+            logger.info("loading vocabulary file {} from cache at {}".format(
+                vocab_file, resolved_vocab_file))
+            logger.info("loading merges file {} from cache at {}".format(
+                merges_file, resolved_merges_file))
+        if pretrained_model_name_or_path in PRETRAINED_VOCAB_POSITIONAL_EMBEDDINGS_SIZE_MAP:
+            # if we're using a pretrained model, ensure the tokenizer wont index sequences longer
+            # than the number of positional embeddings
+            max_len = PRETRAINED_VOCAB_POSITIONAL_EMBEDDINGS_SIZE_MAP[pretrained_model_name_or_path]
+            kwargs['max_len'] = min(kwargs.get('max_len', int(1e12)), max_len)
+        # Instantiate tokenizer.
+        if special_tokens_file and 'special_tokens' not in kwargs:
+            special_tokens = open(special_tokens_file, encoding='utf-8').read().split('\n')[:-1]
+        else:
+            special_tokens = kwargs.pop('special_tokens', [])
+        tokenizer = cls(
+            resolved_vocab_file,
+            resolved_merges_file,
+            special_tokens=special_tokens,
+            *inputs,
+            **kwargs)
+        return tokenizer
+
+
+    def __init__(self, vocab_file, merges_file, errors='replace',
+                 special_tokens=None, max_len=None, max_token_len_cache=9):
+        """
+        max_token_len_cache determines whether a normalized token will be cached. It tries to only store shorter tokens in the cache, 
+        with the heuristic that they are more frequent. Increasing this may make tokenization faster but will also take more memory. 
+        The upper bound of the normalized token cache is fixed at 1_000_000 tokens.
+        """
+        self.max_len = max_len if max_len is not None else int(1e12)
+        self.encoder = json.load(open(vocab_file))
+        self.decoder = {v: k for k, v in self.encoder.items()}
+        self.errors = errors  # how to handle errors in decoding
+        self.byte_encoder = bytes_to_unicode()
+        self.byte_decoder = {v: k for k, v in self.byte_encoder.items()}
+        bpe_data = open(merges_file, encoding='utf-8').read().split('\n')[1:-1]
+        bpe_merges = [tuple(merge.split()) for merge in bpe_data]
+        self.bpe_ranks = dict(zip(bpe_merges, range(len(bpe_merges))))
+        # Should haved added re.IGNORECASE so BPE merges can happen for
+        # capitalized versions of contractions
+        self.pat = re.compile(
+            r"""'s|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+""")
+
+        self.special_tokens = {}
+        self.special_tokens_decoder = {}
+        self.set_special_tokens(special_tokens)
+        self.max_token_len_cache = max_token_len_cache
+
+    def __len__(self):
+        return len(self.encoder) + len(self.special_tokens)
+
+    def set_special_tokens(self, special_tokens):
+        """ Add a list of additional tokens to the encoder.
+            The additional tokens are indexed starting from the last index of the
+            current vocabulary in the order of the `special_tokens` list.
+        """
+        if not special_tokens:
+            self.special_tokens = {}
+            self.special_tokens_decoder = {}
+            return
+        self.special_tokens = dict((tok, len(self.encoder) + i)
+                                   for i, tok in enumerate(special_tokens))
+        self.special_tokens_decoder = {v: k for k, v in self.special_tokens.items()}
+        logger.info("Special tokens {}".format(self.special_tokens))
+    
+    @lru_cache(1_000_000)
+    def bpe(self, token):
+        word = tuple(token)
+        pairs = get_pairs(word)
+
+        if not pairs:
+            return token
+
+        while True:
+            bigram = min(pairs, key=lambda pair: self.bpe_ranks.get(pair, float('inf')))
+            if bigram not in self.bpe_ranks:
+                break
+            first, second = bigram
+            new_word = []
+            i = 0
+            while i < len(word):
+                try:
+                    j = word.index(first, i)
+                    new_word.extend(word[i:j])
+                    i = j
+                except BaseException:
+                    new_word.extend(word[i:])
+                    break
+
+                if word[i] == first and i < len(word) - 1 and word[i + 1] == second:
+                    new_word.append(first + second)
+                    i += 2
+                else:
+                    new_word.append(word[i])
+                    i += 1
+            new_word = tuple(new_word)
+            word = new_word
+            if len(word) == 1:
+                break
+            else:
+                pairs = get_pairs(word)
+        word = ' '.join(word)
+        return word
+      
+    @lru_cache(1_000_000)
+    def normalize_token_and_cache(self, token):
+        return self.normalize_token(token)
+    
+    def normalize_token(self, token):
+          token = ''.join(self.byte_encoder[b] for b in token.encode('utf-8'))
+          ret = [bpe_token for bpe_token in self.bpe(token).split(' ')]
+          return ret
+
+    def tokenize(self, text):
+        """ Tokenize a string. """
+        max_token_len_cache = self.max_token_len_cache
+        bpe_tokens = []
+        if sys.version_info[0] == 2:
+          for token in re.findall(self.pat, text):        
+              token = ''.join(self.byte_encoder[ord(b)] for b in token)
+              bpe_tokens.extend(bpe_token for bpe_token in self.bpe(token).split(' '))
+          return bpe_tokens
+        for token in re.findall(self.pat, text):
+          if len(token) <= max_token_len_cache:
+              bpe_tokens.extend(self.normalize_token_and_cache(token))
+          else: 
+              bpe_tokens.extend(self.normalize_token(token))
+        return bpe_tokens
+      
+    def convert_tokens_to_ids(self, tokens):
+        """ Converts a sequence of tokens into ids using the vocab. """
+        ids = []
+        if isinstance(tokens, str) or (sys.version_info[0] == 2 and isinstance(tokens, unicode)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.encoder.get(tokens, 0)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.encoder.get(token, 0))
+        if len(ids) > self.max_len:
+            logger.warning(
+                "Token indices sequence length is longer than the specified maximum "
+                " sequence length for this OpenAI GPT model ({} > {}). Running this"
+                " sequence through the model will result in indexing errors".format(
+                    len(ids), self.max_len)
+            )
+        return ids
+
+    def convert_ids_to_tokens(self, ids, skip_special_tokens=False):
+        """Converts a sequence of ids in BPE tokens using the vocab."""
+        tokens = []
+        for i in ids:
+            if i in self.special_tokens_decoder:
+                if not skip_special_tokens:
+                    tokens.append(self.special_tokens_decoder[i])
+            else:
+                tokens.append(self.decoder[i])
+        return tokens
+
+    def encode(self, text):
+        return self.convert_tokens_to_ids(self.tokenize(text))
+
+    def decode(self, tokens):
+        text = ''.join([self.decoder[token] for token in tokens])
+        text = bytearray([self.byte_decoder[c] for c in text]).decode('utf-8', errors=self.errors)
+        return text
+
+    def save_vocabulary(self, vocab_path):
+        """Save the tokenizer vocabulary and merge files to a directory."""
+        if not os.path.isdir(vocab_path):
+            logger.error("Vocabulary path ({}) should be a directory".format(vocab_path))
+            return
+        vocab_file = os.path.join(vocab_path, VOCAB_NAME)
+        merge_file = os.path.join(vocab_path, MERGES_NAME)
+        special_tokens_file = os.path.join(vocab_path, SPECIAL_TOKENS_NAME)
+
+        with open(vocab_file, 'w', encoding='utf-8') as f:
+            f.write(json.dumps(self.encoder, ensure_ascii=False))
+
+        index = 0
+        with open(merge_file, "w", encoding="utf-8") as writer:
+            writer.write(u'#version: 0.2\n')
+            for bpe_tokens, token_index in sorted(self.bpe_ranks.items(), key=lambda kv: kv[1]):
+                if index != token_index:
+                    logger.warning("Saving vocabulary to {}: BPE merge indices are not consecutive."
+                                   " Please check that the tokenizer is not corrupted!".format(merge_file))
+                    index = token_index
+                writer.write(' '.join(bpe_tokens) + u'\n')
+                index += 1
+
+        index = len(self.encoder)
+        with open(special_tokens_file, 'w', encoding='utf-8') as writer:
+            for token, token_index in sorted(self.special_tokens.items(), key=lambda kv: kv[1]):
+                if index != token_index:
+                    logger.warning("Saving special tokens vocabulary to {}: BPE indices are not consecutive."
+                                   " Please check that the tokenizer is not corrupted!".format(special_tokens_file))
+                    index = token_index
+                writer.write(token + u'\n')
+                index += 1
+
+        return vocab_file, merge_file, special_tokens_file

megatron/tokenizer/tokenizer.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Megatron tokenizers."""
+
+from abc import ABC
+from abc import abstractmethod
+from transformers import AutoTokenizer
+
+from .bert_tokenization import FullTokenizer as FullBertTokenizer
+from .gpt2_tokenization import GPT2Tokenizer
+
+def build_tokenizer(args):
+    """Initialize tokenizer."""
+    if args.rank == 0:
+        print('> building {} tokenizer ...'.format(args.tokenizer_type),
+              flush=True)
+
+    # Select and instantiate the tokenizer.
+    assert args.vocab_file is not None or args.tokenizer_type == "PretrainedFromHF"
+    if args.tokenizer_type == 'BertWordPieceLowerCase':
+        tokenizer = _BertWordPieceTokenizer(vocab_file=args.vocab_file,
+                                            lower_case=True,
+                                            vocab_extra_ids=args.vocab_extra_ids)
+    elif args.tokenizer_type == 'BertWordPieceCase':
+        tokenizer = _BertWordPieceTokenizer(vocab_file=args.vocab_file,
+                                            lower_case=False,
+                                            vocab_extra_ids=args.vocab_extra_ids)
+    elif args.tokenizer_type == 'GPT2BPETokenizer':
+        assert args.merge_file is not None
+        tokenizer = _GPT2BPETokenizer(args.vocab_file, args.merge_file)
+    elif args.tokenizer_type == "PretrainedFromHF":
+        assert args.tokenizer_name_or_path is not None
+
+        # prevent transformers from logging info and warnings on each rank
+        import transformers
+        import logging
+        if args.rank == 0:
+            transformers.utils.logging.set_verbosity(logging.INFO)
+        else:
+            # shut the warnings on replicas
+            transformers.utils.logging.set_verbosity(logging.ERROR)
+
+        if args.rank == 0:
+            print(" vocab file is un-used. loading tokenizer from pre-trained model")
+        tokenizer = _AutoTokenizer(args.tokenizer_name_or_path, vocab_extra_ids=args.vocab_extra_ids)
+    else:
+        raise NotImplementedError('{} tokenizer is not '
+                                  'implemented.'.format(args.tokenizer_type))
+
+    # Add vocab size.
+    args.padded_vocab_size = _vocab_size_with_padding(tokenizer.vocab_size,
+                                                      args)
+
+    return tokenizer
+
+
+def _vocab_size_with_padding(orig_vocab_size, args):
+    """Apply the requested rules to change the size of the vocabulary"""
+    if args.pad_vocab_size_to is not None:
+        if args.pad_vocab_size_to  < orig_vocab_size:
+            raise ValueError(
+                f"You asked to pad the vocabulary to {args.pad_vocab_size_to} when the initial vocabulary size is "
+                f"{orig_vocab_size}. You can only pad to a higher value."
+            )
+
+        if args.make_vocab_size_divisible_by is not None and (args.pad_vocab_size_to % args.make_vocab_size_divisible_by) != 0:
+            raise ValueError(f"{args.pad_vocab_size_to} is not divisible by {args.make_vocab_size_divisible_by}")
+
+        after = args.pad_vocab_size_to
+    else:
+        # Pad vocab size so it is divisible by model parallel size and still having GPU friendly size.
+        after = orig_vocab_size
+        multiple = args.make_vocab_size_divisible_by * \
+            args.tensor_model_parallel_size
+        while (after % multiple) != 0:
+            after += 1
+    if args.rank == 0:
+        print(' > padded vocab (size: {}) with {} dummy tokens '
+              '(new size: {})'.format(
+                  orig_vocab_size, after - orig_vocab_size, after), flush=True)
+    return after
+
+
+class AbstractTokenizer(ABC):
+    """Abstract class for tokenizer."""
+
+    def __init__(self, name):
+        self.name = name
+        super().__init__()
+
+    @property
+    @abstractmethod
+    def vocab_size(self):
+        pass
+
+    @property
+    @abstractmethod
+    def vocab(self):
+        """Dictionary from vocab text token to id token."""
+        pass
+
+    @property
+    @abstractmethod
+    def inv_vocab(self):
+        """Dictionary from vocab id token to text token."""
+        pass
+
+    @abstractmethod
+    def tokenize(self, text):
+        pass
+
+    def detokenize(self, token_ids):
+        raise NotImplementedError('detokenizer is not implemented for {} '
+                                  'tokenizer'.format(self.name))
+
+    @property
+    def cls(self):
+        raise NotImplementedError('CLS is not provided for {} '
+                                  'tokenizer'.format(self.name))
+
+    @property
+    def sep(self):
+        raise NotImplementedError('SEP is not provided for {} '
+                                  'tokenizer'.format(self.name))
+
+    @property
+    def pad(self):
+        raise NotImplementedError('PAD is not provided for {} '
+                                  'tokenizer'.format(self.name))
+
+    @property
+    def eod(self):
+        raise NotImplementedError('EOD is not provided for {} '
+                                  'tokenizer'.format(self.name))
+
+    @property
+    def mask(self):
+        raise NotImplementedError('MASK is not provided for {} '
+                                  'tokenizer'.format(self.name))
+
+
+class _BertWordPieceTokenizer(AbstractTokenizer):
+    """Original BERT wordpiece tokenizer."""
+
+    def __init__(self, vocab_file, lower_case=True, vocab_extra_ids=0):
+        if lower_case:
+            name = 'BERT Lower Case'
+        else:
+            name = 'BERT Upper Case'
+        super().__init__(name)
+        self.tokenizer = FullBertTokenizer(vocab_file, do_lower_case=lower_case)
+        self.cls_id = self.tokenizer.vocab['[CLS]']
+        self.sep_id = self.tokenizer.vocab['[SEP]']
+        self.pad_id = self.tokenizer.vocab['[PAD]']
+        self.mask_id = self.tokenizer.vocab['[MASK]']
+        self._additional_special_tokens = []
+
+        # (dsachan) Add BOS and EOS tokens
+        SPECIAL_TOKENS = {'eos_token': '[EOS]',
+                          'bos_token': '[BOS]'}
+        self._bos_token = '[BOS]'
+        self.add_token(self._bos_token)
+        self._bos_token_id = self.vocab.get(self._bos_token)
+
+        self._eos_token = '[EOS]'
+        self.add_token(self._eos_token)
+        self._eos_token_id = self.vocab.get(self._eos_token)
+
+        # (dsachan) Add additional special tokens
+        # These can be used as sentinel tokens in T5 model inputs
+        additional_special_tokens = []
+        additional_special_tokens.extend(
+            ["<extra_id_{}>".format(i) for i in range(vocab_extra_ids)])
+        self.add_additional_special_tokens(additional_special_tokens)
+
+    def add_token(self, token):
+        if token not in self.vocab:
+            self.inv_vocab[self.vocab_size] = token
+            # self.vocab_size comes from len(vocab)
+            # and it will increase as we add elements
+            self.vocab[token] = self.vocab_size
+
+    def add_additional_special_tokens(self, tokens_list):
+        setattr(self, "additional_special_tokens", tokens_list)
+        for value in tokens_list:
+            self.add_token(value)
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.vocab_size()
+
+    @property
+    def vocab(self):
+        return self.tokenizer.vocab
+
+    @property
+    def inv_vocab(self):
+        return self.tokenizer.inv_vocab
+
+    def tokenize(self, text):
+        text_tokens = self.tokenizer.tokenize(text)
+        return self.tokenizer.convert_tokens_to_ids(text_tokens)
+
+    def decode(self, ids):
+        tokens = self.tokenizer.convert_ids_to_tokens(ids)
+        return self.tokenizer.convert_tokens_to_string(tokens)
+
+    def decode_token_ids(self, token_ids):
+        tokens = self.tokenizer.convert_ids_to_tokens(token_ids)
+        exclude_list = ['[PAD]', '[CLS]']
+        non_pads = [t for t in tokens if t not in exclude_list]
+
+        result = ""
+        for s in non_pads:
+            if s.startswith("##"):
+                result += s[2:]
+            else:
+                result += " " + s
+
+        return result
+
+    @property
+    def cls(self):
+        return self.cls_id
+
+    @property
+    def sep(self):
+        return self.sep_id
+
+    @property
+    def pad(self):
+        return self.pad_id
+
+    @property
+    def mask(self):
+        return self.mask_id
+
+    @property
+    def bos_token(self):
+        """ Beginning of sentence token id """
+        return self._bos_token
+
+    @property
+    def eos_token(self):
+        """ End of sentence token id """
+        return self._eos_token
+
+    @property
+    def additional_special_tokens(self):
+        """ All the additional special tokens you may want to use (list of strings)."""
+        return self._additional_special_tokens
+
+    @property
+    def bos_token_id(self):
+        """ Id of the beginning of sentence token in the vocabulary."""
+        return self._bos_token_id
+
+    @property
+    def eos_token_id(self):
+        """ Id of the end of sentence token in the vocabulary."""
+        return self._eos_token_id
+
+    @property
+    def additional_special_tokens_ids(self):
+        """ Ids of all the additional special tokens in the vocabulary (list of integers)."""
+        return [self.vocab.get(token) for token in self._additional_special_tokens]
+
+    @additional_special_tokens.setter
+    def additional_special_tokens(self, value):
+        self._additional_special_tokens = value
+
+
+class _GPT2BPETokenizer(AbstractTokenizer):
+    """Original GPT2 BPE tokenizer."""
+
+    def __init__(self, vocab_file, merge_file):
+        name = 'GPT2 BPE'
+        super().__init__(name)
+
+        self.tokenizer = GPT2Tokenizer(vocab_file, merge_file, errors='replace',
+                                       special_tokens=[], max_len=None)
+        self.eod_id = self.tokenizer.encoder['<|endoftext|>']
+
+    @property
+    def vocab_size(self):
+        return len(self.tokenizer.encoder)
+
+    @property
+    def vocab(self):
+        return self.tokenizer.encoder
+
+    @property
+    def inv_vocab(self):
+        return self.tokenizer.decoder
+
+    def tokenize(self, text):
+        return self.tokenizer.encode(text)
+
+    def detokenize(self, token_ids):
+        return self.tokenizer.decode(token_ids)
+
+    @property
+    def eod(self):
+        return self.eod_id
+
+
+class _AutoTokenizer(AbstractTokenizer):
+    """AutoTokenizer for Hf Pretrained model loading."""
+
+    def __init__(self, tokenizer_name_or_path, vocab_extra_ids):
+        name = tokenizer_name_or_path
+        super().__init__(name)
+        hf_tokenizer_kwargs = {}
+        if vocab_extra_ids > 0:
+            # TODO @thomasw21 we might need to concatenate to a pre-existing list?
+            hf_tokenizer_kwargs["additional_special_tokens"] = [f"<extra_id_{_id}>" for _id in range(vocab_extra_ids)]
+        self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name_or_path, **hf_tokenizer_kwargs)
+        self.encoder = self.tokenizer.get_vocab()
+        self.decoder = {v: k for k, v in self.encoder.items()}
+
+    @property
+    def vocab_size(self):
+        return len(self.tokenizer) # vocab_size doesn't contain additional tokens
+
+    @property
+    def vocab(self):
+        # TODO @thomasw21 make sure that special tokens don't collapse with vocab tokens.
+        return {
+            **{special_token: self.tokenizer.convert_tokens_to_ids(special_token) for special_token in self.tokenizer.additional_special_tokens},
+            **self.tokenizer.vocab,
+        }
+
+    @property
+    def inv_vocab(self):
+        return {v: k for k, v in self.vocab.items()}
+
+    def tokenize(self, text):
+        return self.tokenizer.encode(text)
+
+    def detokenize(self, token_ids):
+        return self.tokenizer.decode(token_ids)
+
+    @property
+    def eod(self):
+        # TODO @thomasw21 might conflict with <eos>
+        return self.eos
+
+    @property
+    def cls(self):
+        candidate = self.tokenizer.cls_token_id
+        return self._check_token_candidate(candidate)
+
+    @property
+    def sep(self):
+        candidate = self.tokenizer.sep_token_id
+        return self._check_token_candidate(candidate)
+
+    @property
+    def pad(self):
+        candidate = self.tokenizer.pad_token_id
+        return self._check_token_candidate(candidate)
+
+    @property
+    def mask(self):
+        candidate = self.tokenizer.mask_token_id
+        return self._check_token_candidate(candidate)
+
+    @property
+    def bos(self):
+        raise NotImplementedError("Missing <bos>")
+
+    @property
+    def eos(self):
+        # TODO @thomasw21 might conflict with the notion of <eod>
+        candidate = self.tokenizer.eos_token_id
+        return self._check_token_candidate(candidate)
+
+    @property
+    def additional_special_tokens_ids(self):
+        """ All the additional special tokens you may want to use (list of strings)."""
+        return self.tokenizer.additional_special_tokens_ids
+
+    @staticmethod
+    def _check_token_candidate(candidate):
+        if candidate is None:
+            raise AttributeError("Token doesn't exist")
+        return candidate

megatron/training.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Pretrain utilities."""
+
+from datetime import datetime
+import bisect
+import math
+import sys
+import time
+import json
+# The earliest we can measure the start time.
+_TRAIN_START_TIME = time.time()
+
+import torch
+from torch.nn.parallel.distributed import DistributedDataParallel as torchDDP
+
+from megatron import get_args
+from megatron import get_timers
+from megatron import get_tensorboard_writer
+from megatron import get_current_global_batch_size
+from megatron import get_num_microbatches
+from megatron import is_last_rank
+from megatron import update_num_microbatches
+from megatron import mpu
+from megatron import print_rank_0
+from megatron import print_rank_last
+from megatron.checkpointing import load_checkpoint
+from megatron.checkpointing import save_checkpoint
+from megatron.model.module import Float16Module
+from megatron.optimizer import get_megatron_optimizer
+from megatron.initialize import initialize_megatron
+from megatron.initialize import write_args_to_tensorboard, log_restart_to_tensorboard
+from megatron.learning_rates import AnnealingLR
+from megatron.model.distributed import DistributedDataParallel as LocalDDP
+from megatron.utils import check_adlr_autoresume_termination, get_parameters_in_billions
+from megatron.utils import unwrap_model, found_kill_switch
+from megatron.data.data_samplers import build_pretraining_data_loader
+from megatron.utils import calc_params_l2_norm
+from megatron.schedules import forward_backward_no_pipelining
+from megatron.schedules import forward_backward_pipelining_without_interleaving
+from megatron.schedules import forward_backward_pipelining_with_interleaving
+from megatron.utils import report_memory, flops_calculator
+from megatron.global_vars import codecarbon_tracker_start, codecarbon_tracker_stop
+from megatron.data.dataset_utils import analyze_data_prefix
+
+import deepspeed
+
+
+def print_datetime(string):
+    """Note that this call will sync across all ranks."""
+    torch.distributed.barrier()
+    time_str = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
+    print_rank_0('[' + string + '] datetime: {} '.format(time_str))
+
+
+def pretrain(train_valid_test_dataset_provider,
+             model_provider,
+             forward_step_func,
+             extra_args_provider=None,
+             args_defaults={}):
+    """Main training program.
+
+    This function will run the followings in the order provided:
+        1) initialize Megatron.
+        2) setup model, optimizer and lr schedule using the model_provider.
+        3) call train_val_test_data_provider to get train/val/test datasets.
+        4) train the modle using the forward_step_func.
+
+    Arguments:
+        train_valid_test_dataset_provider: a function that takes the size of
+            train/valid/test dataset and returns `train, valid, test` datasets.
+        model_provider: a function that returns a vanilla version of the
+            model. By vanilla we mean a simple model on cpu with no fp16 or ddp.
+        forward_step_func: a function that takes a `data iterator` and `model`,
+            and returns a `loss` scalar with a dictionary with key:values being
+            the info we would like to monitor during training, for example
+            `lm-loss: value`. We also require that this function add
+            `batch generator` to the timers class.
+        extra_args_provider: a function that takes a parser and adds arguments
+            to it. It is used for programs to add their own arguments.
+        args_defaults: a dictionary from argument-name to argument-value. It
+            to set already parse arguments.
+    """
+
+    # Initalize and get arguments, timers, and Tensorboard writer.
+    initialize_megatron(extra_args_provider=extra_args_provider,
+                        args_defaults=args_defaults)
+
+    args = get_args()
+
+    if found_kill_switch():
+        print_datetime(f"Detected kill switch at {args.kill_switch_path}. Exiting")
+        sys.exit()
+
+    codecarbon_tracker_start()
+
+    # Adjust the startup time so it reflects the largest value.
+    # This will be closer to what scheduler will see (outside of
+    # image ... launches.
+    global _TRAIN_START_TIME
+    start_time_tensor = torch.cuda.FloatTensor([_TRAIN_START_TIME])
+    torch.distributed.all_reduce(start_time_tensor,
+                                 op=torch.distributed.ReduceOp.MIN)
+    _TRAIN_START_TIME = start_time_tensor.item()
+    print_rank_0('time to initialize megatron (seconds): {:.3f}'.format(
+        time.time() - _TRAIN_START_TIME))
+    print_datetime('after megatron is initialized')
+
+    timers = get_timers()
+
+    if args.deepspeed:
+        args.deepspeed_configuration = json.load(
+            open(args.deepspeed_config, 'r', encoding='utf-8'))
+        if "curriculum_learning" in args.deepspeed_configuration and \
+            "enabled" in args.deepspeed_configuration["curriculum_learning"]:
+            args.curriculum_learning = args.deepspeed_configuration[ \
+                "curriculum_learning"]["enabled"]
+        if args.curriculum_learning and \
+            args.pipeline_model_parallel_size >= 1:
+            from deepspeed.runtime.data_pipeline.curriculum_scheduler \
+                import CurriculumScheduler
+            args.curriculum_scheduler = CurriculumScheduler( \
+                args.deepspeed_configuration["curriculum_learning"])
+
+
+    # Model, optimizer, and learning rate.
+    timers('model-and-optimizer-setup').start()
+    model, optimizer, lr_scheduler = setup_model_and_optimizer(model_provider)
+    args.parameters_in_billions_no_embedding = get_parameters_in_billions(model, exclude_embeddings=True)
+    print_rank_0(f'estimated model parameters: {get_parameters_in_billions(model)}')
+    print_rank_0(f'estimated model parameters without embeddings: {get_parameters_in_billions(model, exclude_embeddings=True)}')
+    timers('model-and-optimizer-setup').stop()
+    print_datetime('after model, optimizer, and learning rate '
+                   'scheduler are built')
+
+    # Data stuff.
+    timers('train/valid/test-data-iterators-setup').start()
+    if args.virtual_pipeline_model_parallel_size is not None:
+        all_data_iterators = [
+            build_train_valid_test_data_iterators(train_valid_test_dataset_provider)
+            for _ in range(len(model))
+        ]
+        train_data_iterator = [data_iterators[0] for data_iterators in all_data_iterators]
+        valid_data_iterator = [data_iterators[1] for data_iterators in all_data_iterators]
+        test_data_iterator = [data_iterators[2] for data_iterators in all_data_iterators]
+    else:
+        train_data_iterator, valid_data_iterator, test_data_iterator = build_train_valid_test_data_iterators(
+                train_valid_test_dataset_provider)
+
+    if args.data_path is not None and len(args.data_path) > 1:
+        prefixes, weights = analyze_data_prefix(args.data_path)
+        setattr(args, "data_prefixes", prefixes)
+        setattr(args, "data_weights", weights)
+    elif args.train_weighted_split_paths is not None and len(args.train_weighted_split_paths[0]) > 1:
+        paths = args.train_weighted_split_paths[0]
+        weights = args.train_weighted_split_weights[0]
+        data_prefix = [j for i in [[w,p] for w,p in zip(weights, paths)] for j in i]
+        prefixes, weights = analyze_data_prefix(data_prefix)
+        setattr(args, "data_prefixes", prefixes)
+        setattr(args, "data_weights", weights)
+    else:
+        setattr(args, "data_prefixes", None)
+        setattr(args, "data_weights", None)
+
+    timers('train/valid/test-data-iterators-setup').stop()
+    print_datetime('after dataloaders are built')
+
+    # Print setup timing.
+    print_rank_0('done with setup ...')
+    timers.log(['model-and-optimizer-setup', 'train/valid/test-data-iterators-setup'])
+    print_rank_0('training ...')
+
+    iteration = 0
+    if args.do_train and args.train_iters > 0:
+        iteration = train(forward_step_func,
+                          model, optimizer, lr_scheduler,
+                          train_data_iterator, valid_data_iterator)
+    print_datetime('after training is done')
+
+    if args.do_valid:
+        names = args.valid_weighted_split_names
+        names = names if names is not None else ['valid'] * len(valid_data_iterator)
+        for iterator, name in zip(valid_data_iterator, names):
+            prefix = 'the end of training for val data'
+            evaluate_and_print_results(prefix, forward_step_func,
+                                       iterator, model,
+                                       iteration, False, data_group_name=name)
+
+    if args.save and iteration != 0:
+        save_checkpoint(iteration, model, optimizer, lr_scheduler)
+
+    if args.do_test:
+        # Run on test data.
+        prefix = 'the end of training for test data'
+        names = args.test_weighted_split_names
+        names = names if names is not None else ['test'] * len(test_data_iterator)
+        for iterator, name in zip(test_data_iterator, names):
+            evaluate_and_print_results(prefix, forward_step_func,
+                                       iterator, model,
+                                       0, True, data_group_name=name)
+
+    codecarbon_tracker_stop()
+
+
+def update_train_iters(args):
+
+    # For iteration-based training, we don't need to do anything
+    if args.train_iters:
+        return
+
+    # Constant batch size with sample-based training.
+    if args.rampup_batch_size is None:
+        args.train_iters = args.train_samples // args.global_batch_size
+
+    else:
+        # Sample based training with rampup batch size.
+        iterations = 0
+        consumed_samples = 0
+        # Rampup phase.
+        while consumed_samples <= int(args.rampup_batch_size[2]):
+            update_num_microbatches(consumed_samples, consistency_check=False)
+            consumed_samples += get_current_global_batch_size()
+            iterations += 1
+        # Reset
+        update_num_microbatches(0, consistency_check=False)
+        # Constant phase
+        # Note that we throw away any partial last batch.
+        iterations += (args.train_samples - consumed_samples) // \
+                      args.global_batch_size
+        args.train_iters = iterations
+
+    print_rank_0('setting training iterations to {}'.format(args.train_iters))
+
+
+def get_model(model_provider_func):
+    """Build the model."""
+    args = get_args()
+
+    # Build model.
+    if mpu.get_pipeline_model_parallel_world_size() > 1 and \
+       args.virtual_pipeline_model_parallel_size is not None:
+        model = []
+        for i in range(args.virtual_pipeline_model_parallel_size):
+            mpu.set_virtual_pipeline_model_parallel_rank(i)
+            # Set pre_process and post_process only after virtual rank is set.
+            pre_process = mpu.is_pipeline_first_stage()
+            post_process = mpu.is_pipeline_last_stage()
+            this_model = model_provider_func(
+                pre_process=pre_process,
+                post_process=post_process
+            )
+            model.append(this_model)
+    else:
+        pre_process = mpu.is_pipeline_first_stage()
+        post_process = mpu.is_pipeline_last_stage()
+        model = model_provider_func(
+            pre_process=pre_process,
+            post_process=post_process
+        )
+
+
+    if not isinstance(model, list):
+        model = [model]
+
+    # Set tensor model parallel attributes if not set.
+    # Only parameters that are already tensor model parallel have these
+    # attributes set for them. We should make sure the default attributes
+    # are set for all params so the optimizer can use them.
+    for model_module in model:
+        for param in model_module.parameters():
+            mpu.set_defaults_if_not_set_tensor_model_parallel_attributes(param)
+
+    # # Print number of parameters.
+    # Moved to `train` with extras
+    # if mpu.get_data_parallel_rank() == 0:
+    #     print('Number of parameters on tensor={}, pipeline={}: {}'.format(
+    #         mpu.get_tensor_model_parallel_rank(),
+    #         mpu.get_pipeline_model_parallel_rank(),
+    #         sum([sum([p.ds_numel if hasattr(p,'ds_id') else p.nelement() for p in model_module.parameters()])
+    #              for model_module in model])), flush=True)
+    #     torch.distributed.barrier()
+    # else:
+    #     torch.distributed.barrier()
+
+    if args.deepspeed:
+        return model
+
+    # GPU allocation.
+    for model_module in model:
+        model_module.cuda(torch.cuda.current_device())
+
+    # Fp16 conversion.
+    if args.fp16 or args.bf16:
+        model = [Float16Module(model_module, args) for model_module in model]
+
+    if args.DDP_impl == 'torch':
+        i = torch.cuda.current_device()
+        model = [torchDDP(model_module, device_ids=[i], output_device=i,
+                          process_group=mpu.get_data_parallel_group())
+                 for model_module in model]
+        return model
+
+    if args.DDP_impl == 'local':
+        model = [LocalDDP(model_module,
+                          args.accumulate_allreduce_grads_in_fp32,
+                          args.use_contiguous_buffers_in_ddp)
+                 for model_module in model]
+        return model
+
+    raise NotImplementedError('Unknown DDP implementation specified: {}. '
+                              'Exiting.'.format(args.DDP_impl))
+
+
+def get_learning_rate_scheduler(optimizer):
+    """Build the learning rate scheduler."""
+    args = get_args()
+
+    # Iteration-based training.
+    if args.train_iters:
+        if args.lr_decay_iters is None:
+            args.lr_decay_iters = args.train_iters
+        decay_steps = args.lr_decay_iters * args.global_batch_size
+        if args.lr_warmup_fraction is not None:
+            warmup_steps = args.lr_warmup_fraction * decay_steps
+        else:
+            warmup_steps = args.lr_warmup_iters * args.global_batch_size
+    # Sample-based training.
+    elif args.train_samples:
+        # We need to set training iters for later use. Technically
+        # we need to adjust the training samples too (due to last
+        # batch being incomplete) but we leave it as is for now.
+        update_train_iters(args)
+        if args.lr_decay_samples is None:
+            args.lr_decay_samples = args.train_samples
+        decay_steps = args.lr_decay_samples
+        if args.lr_warmup_fraction is not None:
+            warmup_steps = args.lr_warmup_fraction * decay_steps
+        else:
+            warmup_steps = args.lr_warmup_samples
+    else:
+        raise Exception(
+            'either train-iters or train-samples should be provided.')
+
+    lr_scheduler = AnnealingLR(
+        optimizer,
+        max_lr=args.lr,
+        min_lr=args.min_lr,
+        warmup_steps=warmup_steps,
+        decay_steps=decay_steps,
+        decay_style=args.lr_decay_style,
+        use_checkpoint_lr_scheduler=args.use_checkpoint_lr_scheduler,
+        override_lr_scheduler=args.override_lr_scheduler)
+
+    return lr_scheduler
+
+
+def sync_hp_to_lp(optimizer):
+
+    optimizer.update_lp_params()
+
+    # for n,p in model.named_parameters():
+    #     print(n)
+
+    #     if p._hp_mapping is not None:
+    #         #print(f'rank {rank} fixing hp for input_layernorm')
+    #         #p._hp_mapping.update_hp()
+
+    #         hp = p._hp_mapping.hp_fragment
+
+
+
+    #         torch.distributed.all_reduce(hp, op=torch.distributed.ReduceOp.AVG, group=mpu.get_tensor_model_parallel_group())
+
+    #         # 3. optim states
+    #         for key in ['exp_avg', 'exp_avg_sq']:
+    #             optim_state_fragment = p._hp_mapping.get_optim_state_fragment(key)
+    #             #print(f'rank {rank} before reduce optim state fragment {key} = {optim_state_fragment}')
+    #             torch.distributed.all_reduce(optim_state_fragment, op=torch.distributed.ReduceOp.AVG, group=mpu.get_tensor_model_parallel_group())
+    #             #print(f'rank {rank} after reduce optim state fragment {key} = {optim_state_fragment}')
+
+
+
+def setup_model_and_optimizer(model_provider_func):
+    """Setup model and optimizer."""
+    args = get_args()
+
+    model = get_model(model_provider_func)
+
+    unwrapped_model = unwrap_model(model,
+                                   (torchDDP, LocalDDP, Float16Module))
+
+    if args.inference:
+        optimizer = None
+        lr_scheduler = None
+    else:
+        optimizer = get_megatron_optimizer(unwrapped_model)
+        lr_scheduler = get_learning_rate_scheduler(optimizer)
+
+
+    if args.deepspeed:
+        print_rank_0("DeepSpeed is enabled.")
+        #pp = mpu.get_pipeline_model_parallel_world_size()
+
+        import json
+        import io
+        with io.open(args.deepspeed_config, "r", encoding="utf-8") as f:
+            config = json.load(f)
+        if args.universal_checkpoint:
+            config["checkpoint"] = {"load_universal": True}
+
+        model, optimizer, _, lr_scheduler = deepspeed.initialize(
+            model=model[0],
+            optimizer=optimizer,
+            lr_scheduler=lr_scheduler,
+            # config=config,
+            args=args,
+        )
+
+        assert model.fp16_enabled() == args.fp16, "megatron fp16 config does not match deepspeed"
+        assert model.bfloat16_enabled() == args.bf16, "megatron bf16 config does not match deepspeed"
+
+        if isinstance(model, deepspeed.PipelineEngine):
+            # hack to get batch_fn from pretrain_gpt.py
+            model.set_batch_fn(model.module._megatron_batch_fn)
+
+            assert model.grid.get_pipe_parallel_rank() == mpu.get_pipeline_model_parallel_rank()
+            assert model.grid.get_slice_parallel_rank() == mpu.get_tensor_model_parallel_rank()
+            assert model.grid.get_data_parallel_rank() == mpu.get_data_parallel_rank()
+        model = [model]
+
+    if args.load is not None:
+        timers = get_timers()
+        # Extra barrier is added to make sure all ranks report the
+        # max time.
+        torch.distributed.barrier()
+        timers('load-checkpoint').start()
+        args.iteration = load_checkpoint(model, optimizer, lr_scheduler)
+        torch.distributed.barrier()
+        timers('load-checkpoint').stop()
+        timers.log(['load-checkpoint'])
+
+
+        # hp -> lp
+        if args.deepspeed and args.universal_checkpoint:
+            sync_hp_to_lp(optimizer)
+
+
+    else:
+        args.iteration = 0
+
+    # tp_rank = mpu.get_tensor_model_parallel_rank()
+    # pp_rank = mpu.get_pipeline_model_parallel_rank()
+    # dp_rank = mpu.get_data_parallel_rank()
+    # for n,p in model[0].named_parameters():
+    #     if 'word_embeddings.weight' not in n:
+    #         continue 
+    #     if tp_rank == 0 and pp_rank == 0:
+    #         print(f"{tp_rank=}{pp_rank=}{dp_rank=} bf16 {n=} {p[:10]=}")
+    #         if p._hp_mapping is not None:
+    #             hp = p._hp_mapping.hp_fragment
+    #             print(f'{tp_rank=}{pp_rank=}{dp_rank=} fp32 {n=} {hp[:10]=}')
+
+    #     if tp_rank == 0 and pp_rank == mpu.get_pipeline_model_parallel_world_size() - 1:
+    #         print(f"{tp_rank=}{pp_rank=}{dp_rank=} bf16 {n=} {p[:10]=}")
+    #         if p._hp_mapping is not None:
+    #             hp = p._hp_mapping.hp_fragment
+    #             print(f'{tp_rank=}{pp_rank=}{dp_rank=} fp32 {n=} {hp[:10]=}')
+
+
+    # We only support local DDP with multiple micro-batches.
+    if len(model) > 1 or mpu.get_pipeline_model_parallel_world_size() > 1:
+        assert args.DDP_impl == 'local'
+
+    # get model without FP16 and/or TorchDDP wrappers
+    if args.iteration == 0 and len(unwrapped_model) == 1 \
+        and hasattr(unwrapped_model[0], 'init_state_dict_from_bert'):
+        print_rank_0("Initializing ICT from pretrained BERT model")
+        unwrapped_model[0].init_state_dict_from_bert()
+        if args.fp16:
+            optimizer.reload_model_params()
+
+    return model, optimizer, lr_scheduler
+
+
+def train_step(forward_step_func, data_iterator,
+               model, optimizer, lr_scheduler):
+    """Single training step."""
+    args = get_args()
+    timers = get_timers()
+
+    if args.deepspeed:
+        assert isinstance(model[0], deepspeed.PipelineEngine), model
+        loss = model[0].train_batch(data_iter=data_iterator)
+        skipped_iter = 0
+        grad_norm = model[0].get_global_grad_norm()
+        num_zeros_in_grad = 0
+        return {'lm loss' : loss}, skipped_iter, grad_norm, num_zeros_in_grad
+
+    # Set grad to zero.
+    if not args.deepspeed:
+        if args.DDP_impl == 'local' and args.use_contiguous_buffers_in_ddp:
+            for partition in model:
+                partition.zero_grad_buffer()
+        else:
+            optimizer.zero_grad()
+
+    if mpu.get_pipeline_model_parallel_world_size() > 1:
+        if args.virtual_pipeline_model_parallel_size is not None:
+            forward_backward_func = forward_backward_pipelining_with_interleaving
+            assert get_num_microbatches() % args.pipeline_model_parallel_size == 0, \
+                'number of microbatches is not divisible by pipeline-parallel ' \
+                'size when using interleaved schedule'
+        else:
+            forward_backward_func = forward_backward_pipelining_without_interleaving
+    else:
+        forward_backward_func = forward_backward_no_pipelining
+    losses_reduced = forward_backward_func(
+        forward_step_func, data_iterator, model,
+        optimizer, timers, forward_only=False)
+
+    # All-reduce if needed.
+    if not args.deepspeed and args.DDP_impl == 'local':
+        timers('backward-params-all-reduce').start()
+        for model_module in model:
+            model_module.allreduce_gradients()
+        timers('backward-params-all-reduce').stop()
+
+    # All-reduce word_embeddings' grad across first and last stages to ensure
+    # that word_embeddings parameters stay in sync.
+    # This should only run for models that support pipelined model parallelism
+    # (BERT and GPT-2).
+    timers('backward-embedding-all-reduce').start()
+    if not args.deepspeed:
+        if (mpu.is_pipeline_first_stage(ignore_virtual=True) or
+            mpu.is_pipeline_last_stage(ignore_virtual=True)) and \
+                mpu.get_pipeline_model_parallel_world_size() > 1:
+            if mpu.is_pipeline_first_stage(ignore_virtual=True):
+                unwrapped_model = model[0]
+            elif mpu.is_pipeline_last_stage(ignore_virtual=True):
+                unwrapped_model = model[-1]
+            unwrapped_model = unwrap_model(
+                unwrapped_model, (torchDDP, LocalDDP, Float16Module))
+
+            if unwrapped_model.share_word_embeddings:
+                word_embeddings_weight = unwrapped_model.word_embeddings_weight()
+                if args.DDP_impl == 'local':
+                    grad = word_embeddings_weight.main_grad
+                else:
+                    grad = word_embeddings_weight.grad
+                torch.distributed.all_reduce(grad, group=mpu.get_embedding_group())
+    timers('backward-embedding-all-reduce').stop()
+
+    # Update parameters.
+    timers('optimizer').start()
+    if args.deepspeed:
+        increment = get_num_microbatches() * \
+                    args.micro_batch_size * \
+                    args.data_parallel_size
+        model[0].step(lr_kwargs={'increment': increment})
+        update_successful = model[0].was_step_applied()
+    else:
+        update_successful, grad_norm, num_zeros_in_grad = optimizer.step()
+    timers('optimizer').stop()
+
+    # Update learning rate.
+    if args.deepspeed:
+        skipped_iter = 0
+        grad_norm = None
+        num_zeros_in_grad = None
+    else:
+        if update_successful:
+            increment = get_num_microbatches() * \
+                        args.micro_batch_size * \
+                        args.data_parallel_size
+            lr_scheduler.step(increment=increment)
+            skipped_iter = 0
+        else:
+            skipped_iter = 1
+
+        if mpu.is_pipeline_last_stage(ignore_virtual=True):
+            # Average loss across microbatches.
+            loss_reduced = {}
+            for key in losses_reduced[0]:
+                losses_reduced_for_key = [x[key] for x in losses_reduced]
+                loss_reduced[key] = sum(losses_reduced_for_key) / len(losses_reduced_for_key)
+            return loss_reduced, skipped_iter, grad_norm, num_zeros_in_grad
+    return {}, skipped_iter, grad_norm, num_zeros_in_grad
+
+
+def training_log(loss_dict, total_loss_dict, learning_rate, iteration,
+                 loss_scale, report_memory_flag, skipped_iter,
+                 grad_norm, params_norm, num_zeros_in_grad,
+                 model=None):
+    """Log training information such as losses, timing, ...."""
+    args = get_args()
+    timers = get_timers()
+    writer = get_tensorboard_writer()
+
+    # Advanced, skipped, and Nan iterations.
+    advanced_iters_key = 'advanced iterations'
+    skipped_iters_key = 'skipped iterations'
+    nan_iters_key = 'nan iterations'
+    # Advanced iterations.
+    if not skipped_iter:
+        total_loss_dict[advanced_iters_key] = total_loss_dict.get(
+            advanced_iters_key, 0) + 1
+    else:
+        if advanced_iters_key not in total_loss_dict:
+            total_loss_dict[advanced_iters_key] = 0
+    # Skipped iterations.
+    total_loss_dict[skipped_iters_key] = total_loss_dict.get(
+        skipped_iters_key, 0) + skipped_iter
+    # Update losses and set nan iterations
+    got_nan = False
+    for key in loss_dict:
+        if not skipped_iter:
+            total_loss_dict[key] = total_loss_dict.get(
+                key, torch.cuda.FloatTensor([0.0])) + loss_dict[key]
+        else:
+            value = loss_dict[key].float().sum().item()
+            is_nan = value == float('inf') or \
+                     value == -float('inf') or \
+                     value != value
+            got_nan = got_nan or is_nan
+    total_loss_dict[nan_iters_key] = total_loss_dict.get(
+        nan_iters_key, 0) + int(got_nan)
+
+    # Logging.
+    timers_to_log = []
+
+    def add_to_logging(name):
+        if name in timers.timers:
+            timers_to_log.append(name)
+    add_to_logging('forward-compute')
+    add_to_logging('forward-recv')
+    add_to_logging('forward-send')
+    add_to_logging('forward-backward-send-forward-backward-recv')
+    add_to_logging('backward-compute')
+    add_to_logging('backward-recv')
+    add_to_logging('backward-send')
+    add_to_logging('backward-send-forward-recv')
+    add_to_logging('backward-send-backward-recv')
+    add_to_logging('backward-params-all-reduce')
+    add_to_logging('backward-embedding-all-reduce')
+    add_to_logging('optimizer-copy-to-main-grad')
+    add_to_logging('optimizer-unscale-and-check-inf')
+    add_to_logging('optimizer-clip-main-grad')
+    add_to_logging('optimizer-copy-main-to-model-params')
+    add_to_logging('optimizer')
+    add_to_logging('batch-generator')
+
+    # Calculate batch size.
+    batch_size = args.micro_batch_size * args.data_parallel_size * \
+        get_num_microbatches()
+
+    total_iterations = total_loss_dict[advanced_iters_key] + \
+                       total_loss_dict[skipped_iters_key]
+
+    # Tensorboard values.
+    if writer and (iteration % args.tensorboard_log_interval == 0) and \
+       is_last_rank():
+        writer.add_scalar('steps-vs-samples/y=steps,x=samples', iteration, args.consumed_train_samples)
+        writer.add_scalar('steps-vs-samples/y=samples,x=steps', args.consumed_train_samples, iteration)
+        writer.add_scalar('steps-vs-tokens/y=steps,x=tokens', iteration, args.consumed_train_tokens)
+        writer.add_scalar('steps-vs-tokens/y=tokens,x=steps', args.consumed_train_tokens, iteration)
+        if args.log_learning_rate_to_tensorboard:
+            writer.add_scalar('learning-rate/learning-rate', learning_rate, iteration)
+            writer.add_scalar('learning-rate/learning-rate vs samples', learning_rate,
+                              args.consumed_train_samples)
+            writer.add_scalar('learning-rate/learning-rate vs tokens', learning_rate,
+                              args.consumed_train_tokens)
+        if args.log_batch_size_to_tensorboard:
+            writer.add_scalar('batch-size/batch-size', batch_size, iteration)
+            writer.add_scalar('batch-size/batch-size vs samples', batch_size,
+                              args.consumed_train_samples)
+        for key in loss_dict:
+            writer.add_scalar(f"lm-loss-training/{key}", loss_dict[key], iteration)
+            writer.add_scalar(f"lm-loss-training/{key}" + ' vs samples', loss_dict[key],
+                              args.consumed_train_samples)
+            writer.add_scalar(f"lm-loss-training/{key}" + ' vs tokens', loss_dict[key],
+                              args.consumed_train_tokens)
+
+            writer.add_scalar(f"lm-loss-training/{key}" + ' vs gigaflos (without embeddings)', loss_dict[key],
+                              args.gigaflos_no_embeds)
+        if args.log_loss_scale_to_tensorboard and args.fp16:
+            writer.add_scalar('loss-scale/loss-scale', loss_scale, iteration)
+            writer.add_scalar('loss-scale/loss-scale vs samples', loss_scale,
+                              args.consumed_train_samples)
+            writer.add_scalar('loss-scale/loss-scale vs tokens', loss_scale,
+                              args.consumed_train_tokens)
+        if grad_norm is not None:
+            writer.add_scalar('grad-norm/grad-norm', grad_norm, iteration)
+            writer.add_scalar('grad-norm/grad-norm vs samples', grad_norm,
+                              args.consumed_train_samples)
+            writer.add_scalar('grad-norm/grad-norm vs tokens', grad_norm,
+                              args.consumed_train_tokens)
+        if num_zeros_in_grad is not None:
+            writer.add_scalar('num-zeros/num-zeros', num_zeros_in_grad, iteration)
+            writer.add_scalar('num-zeros/num-zeros vs samples', num_zeros_in_grad,
+                              args.consumed_train_samples)
+            writer.add_scalar('num-zeros/num-zeros vs tokens', num_zeros_in_grad,
+                              args.consumed_train_tokens)
+        if params_norm is not None:
+            writer.add_scalar('params-norm/params-norm', params_norm, iteration)
+            writer.add_scalar('params-norm/params-norm vs samples', params_norm,
+                              args.consumed_train_samples)
+            writer.add_scalar('params-norm/params-norm vs tokens', params_norm,
+                              args.consumed_train_tokens)
+        if args.curriculum_learning:
+            writer.add_scalar('curriculum_seqlen', args.curriculum_seqlen,
+                              iteration)
+
+        # It's very questionable what this data contributes, other than huge unstripped file paths
+        # as keys and hundreds of TB boards that make the TB files very bloated. So disabling for now.
+        #
+        # if args.data_weights is not None:
+        #     for prefix, weight in zip(args.data_prefixes, args.data_weights):
+        #         name = prefix.split(",")[-1]
+        #         writer.add_scalar(f'samples-per-dataset/{name}', args.consumed_train_samples * weight, args.consumed_train_samples)
+        #         writer.add_scalar(f'steps-per-dataset/{name}', iteration * weight, iteration)
+        #         writer.add_scalar(f'tokens-per-dataset/{name}', args.consumed_train_tokens * weight, args.consumed_train_tokens)
+
+        if args.log_timers_to_tensorboard:
+            timers.write(timers_to_log, writer, iteration,
+                         normalizer=total_iterations)
+
+    if iteration % args.log_interval == 0:
+        elapsed_time = timers('interval-time').elapsed()
+        elapsed_time_per_iteration = elapsed_time / total_iterations
+
+        seq_len = args.curriculum_seqlen if args.curriculum_learning else args.seq_length
+        hidden_size = args.hidden_size
+        num_layers = args.num_layers
+        vocab_size = args.padded_vocab_size
+
+        # Compute throughput.
+        samples_per_sec = batch_size / elapsed_time_per_iteration
+        samples_per_sec_per_replica = samples_per_sec / args.data_parallel_size
+        tokens_per_sec = samples_per_sec * seq_len
+        tokens_per_sec_per_replica = tokens_per_sec / args.data_parallel_size
+
+        # General TFLOPs formula (borrowed from Equation 3 in Section 5.1 of
+        # https://arxiv.org/pdf/2104.04473.pdf).
+        # The factor of 4 is when used with activation check-pointing,
+        # otherwise it will be 3, but for 200B model, activation check-pointing will always be on.
+        checkpoint_activations_factor = 4 if args.checkpoint_activations else 3
+        # GLU activations double the hidden states in the upscaling feed-forward in each transformer layer
+        # This leads to 16bsh^2 instead of 8bsh^2 per first feed-forward layer in MLP, thus we increase the coefficient by 8.
+        # Refer to https://github.com/bigscience-workshop/Megatron-DeepSpeed/pull/283#issue-1260805063 for more details.
+        coefficient = 32 if args.glu_activation else 24
+        flops_per_iteration = (coefficient * checkpoint_activations_factor * batch_size * seq_len * num_layers * (hidden_size**2)) * (1. + (seq_len / (6. * hidden_size)) + (vocab_size / (16. * num_layers * hidden_size)))
+        tflops = flops_per_iteration / (elapsed_time_per_iteration * args.world_size * (10**12))
+
+        # only the last rank process has a non-None _GLOBAL_TENSORBOARD_WRITER
+        if writer and is_last_rank():
+            if args.log_timers_to_tensorboard:
+                writer.add_scalar('iteration-time/iteration-time',
+                                  elapsed_time_per_iteration, iteration)
+                writer.add_scalar('iteration-time/iteration-time vs samples',
+                                  elapsed_time_per_iteration, args.consumed_train_samples)
+                writer.add_scalar('iteration-time/iteration-time vs tokens',
+                                  elapsed_time_per_iteration, args.consumed_train_tokens)
+                writer.add_scalar('iteration-time/samples per second',
+                                  samples_per_sec, args.iteration)
+                writer.add_scalar('iteration-time/samples per second per replica',
+                                  samples_per_sec_per_replica, args.iteration)
+                writer.add_scalar('iteration-time/tokens per second',
+                                  tokens_per_sec, args.iteration)
+                writer.add_scalar('iteration-time/tokens per second per replica',
+                                  tokens_per_sec_per_replica, args.iteration)
+                writer.add_scalar('iteration-time/TFLOPs per gpu (estimated)',
+                                  tflops, args.iteration)
+
+        log_string = ' iteration {:8d}/{:8d} |'.format(
+            iteration, args.train_iters)
+        log_string += ' consumed samples: {:12d} |'.format(
+            args.consumed_train_samples)
+        log_string += ' consumed tokens: {:12d} |'.format(
+            args.consumed_train_tokens)
+        log_string += ' elapsed time per iteration (s): {:.2f} |'.format(
+            elapsed_time_per_iteration)
+        log_string += ' learning rate: {:.3E} |'.format(learning_rate)
+        log_string += ' global batch size: {:5d} |'.format(batch_size)
+        for key in total_loss_dict:
+            if key not in [advanced_iters_key, skipped_iters_key,
+                           nan_iters_key]:
+                avg = total_loss_dict[key].item() / \
+                      float(max(1, total_loss_dict[advanced_iters_key]))
+                if avg > 0.0:
+                    log_string += ' {}: {:.6E} |'.format(key, avg)
+                total_loss_dict[key] = torch.cuda.FloatTensor([0.0])
+        if args.fp16:
+            log_string += ' loss scale: {:.1f} |'.format(loss_scale)
+        if grad_norm is not None:
+            log_string += ' grad norm: {:.3f} |'.format(grad_norm)
+        if num_zeros_in_grad is not None:
+            log_string += ' num zeros: {:.1f} |'.format(num_zeros_in_grad)
+        if params_norm is not None:
+            log_string += ' params norm: {:.3f} |'.format(params_norm)
+        if args.curriculum_learning:
+            log_string += ' curriculum seqlen: {:5d} |'.format(args.curriculum_seqlen)
+        log_string += ' number of skipped iterations: {:3d} |'.format(
+            total_loss_dict[skipped_iters_key])
+        log_string += ' number of nan iterations: {:3d} |'.format(
+            total_loss_dict[nan_iters_key])
+        log_string += ' samples per second: {:.3f} |'.format(samples_per_sec)
+        log_string += ' TFLOPs: {:.2f} |'.format(tflops)
+        total_loss_dict[advanced_iters_key] = 0
+        total_loss_dict[skipped_iters_key] = 0
+        total_loss_dict[nan_iters_key] = 0
+        print_rank_last(log_string)
+        if report_memory_flag and learning_rate > 0.:
+            # Report memory after optimizer state has been initialized.
+            report_memory('(after {} iterations)'.format(iteration))
+            report_memory_flag = False
+        timers.log(timers_to_log, normalizer=args.log_interval)
+        flops_calculator(model, args, elapsed_time)
+
+    return report_memory_flag
+
+
+def save_checkpoint_and_time(iteration, model, optimizer, lr_scheduler):
+    timers = get_timers()
+    # Extra barrier is added to make sure
+    # all ranks report the max time.
+    torch.distributed.barrier()
+    timers('save-checkpoint').start()
+    save_checkpoint(iteration, model, optimizer, lr_scheduler)
+    torch.distributed.barrier()
+    timers('save-checkpoint').stop()
+    timers.log(['save-checkpoint'])
+
+
+def train(forward_step_func, model, optimizer, lr_scheduler,
+          train_data_iterator, valid_data_iterator):
+    """Train the model function."""
+    args = get_args()
+    timers = get_timers()
+
+    if args.rank == 0:
+        print("Number of parameters: [tensor rank - pipeline rank] w/ and w/o embeddings:")
+    torch.distributed.barrier()
+    if mpu.get_data_parallel_rank() == 0:
+        tp_rank = mpu.get_tensor_model_parallel_rank()
+        pp_rank = mpu.get_pipeline_model_parallel_rank()
+        preamble = f"[{tp_rank:0>3d}-{pp_rank:0>3d}]"
+        print(f"{preamble} {get_parameters_in_billions(model):.4f}B / {get_parameters_in_billions(model, exclude_embeddings=True):.4f}B", flush=True)
+        torch.distributed.barrier()
+    else:
+        torch.distributed.barrier()
+
+    # Write args to tensorboard
+    write_args_to_tensorboard()
+    log_restart_to_tensorboard()
+
+    # Turn on training mode which enables dropout.
+    for model_module in model:
+        model_module.train()
+
+    # Tracking loss.
+    total_loss_dict = {}
+
+    # Iterations.
+    iteration = args.iteration
+
+    timers('interval-time').start()
+    print_datetime('before the start of training step')
+    report_memory_flag = True
+
+    # flush intervals prior to current iteration
+    if args.skip_train_iteration_range is not None:
+        ends = [end for start, end in args.skip_train_iteration_range]
+        index = bisect.bisect_left(ends, iteration)
+        for _ in range(index):
+            args.skip_train_iteration_range.popleft()
+
+    while iteration < args.train_iters:
+        if (
+            # train_data_iterator is not None
+            args.skip_train_iteration_range is not None
+            and len(args.skip_train_iteration_range) > 0
+            and args.skip_train_iteration_range[0][0] <= iteration + 1 <= args.skip_train_iteration_range[0][1]
+        ):
+            start, end = args.skip_train_iteration_range.popleft()
+            print_rank_0(f"Skipped iterations {start} to {end} due to --skip-train-iteration-range flag.")
+            iteration_for_skipping = args.iteration
+            while iteration_for_skipping + 1 <= end:
+                try:
+                    _ = next(train_data_iterator)
+                except TypeError:
+                    pass
+                iteration_for_skipping += 1
+            continue
+
+        if found_kill_switch():
+            save_checkpoint_and_time(iteration, model, optimizer, lr_scheduler)
+            print_datetime(f"Detected kill switch at {args.kill_switch_path}. Exiting")
+            sys.exit()
+
+        update_num_microbatches(args.consumed_train_samples)
+        if args.deepspeed:
+            # inform deepspeed of any batch size changes
+            global_batch_size = mpu.get_data_parallel_world_size() * \
+                                args.micro_batch_size * \
+                                get_num_microbatches()
+            model[0].set_train_batch_size(global_batch_size)
+
+        if args.curriculum_learning and \
+            args.pipeline_model_parallel_size >= 1:
+            args.curriculum_seqlen = args.curriculum_scheduler.update_difficulty( \
+                    args.iteration + 1)
+        loss_dict, skipped_iter, grad_norm, num_zeros_in_grad = \
+            train_step(forward_step_func,
+                       train_data_iterator,
+                       model,
+                       optimizer,
+                       lr_scheduler)
+        iteration += 1
+        args.iteration = iteration
+        new_samples = mpu.get_data_parallel_world_size() * \
+                                       args.micro_batch_size * \
+                                       get_num_microbatches()
+        args.consumed_train_samples += new_samples
+        if args.curriculum_learning:
+            args.consumed_train_tokens += new_samples * args.curriculum_seqlen
+        else:
+            args.consumed_train_tokens += new_samples * args.seq_length
+        args.gigaflos_no_embeds += (6 * new_samples * args.seq_length * get_parameters_in_billions(model, exclude_embeddings=True))
+
+        # Logging.
+        loss_scale = None
+        if args.fp16:
+            if args.deepspeed:
+                loss_scale = model[0].optimizer.cur_scale
+            else:
+                loss_scale = optimizer.get_loss_scale().item()
+        params_norm = None
+        if args.log_params_norm:
+            params_norm = calc_params_l2_norm(model)
+        report_memory_flag = training_log(loss_dict, total_loss_dict,
+                                          optimizer.param_groups[0]['lr'],
+                                          iteration, loss_scale,
+                                          report_memory_flag, skipped_iter,
+                                          grad_norm, params_norm, num_zeros_in_grad,
+                                          model)
+
+        # Autoresume
+        if args.adlr_autoresume and \
+           (iteration % args.adlr_autoresume_interval == 0):
+            check_adlr_autoresume_termination(iteration, model, optimizer,
+                                              lr_scheduler)
+
+        # Evaluation
+        if args.eval_interval and iteration % args.eval_interval == 0 and \
+           args.do_valid:
+            prefix = 'iteration {}'.format(iteration)
+            names = args.valid_weighted_split_names
+            names = names if names is not None else ['valid'] * len(valid_data_iterator)
+            for iterator, name in zip(valid_data_iterator, names):
+                evaluate_and_print_results(prefix, forward_step_func,
+                                           iterator, model,
+                                           iteration, False, data_group_name=name)
+
+        # Checkpointing
+        saved_checkpoint = False
+        if args.save and args.save_interval and \
+           iteration % args.save_interval == 0:
+            save_checkpoint_and_time(iteration, model, optimizer,
+                                     lr_scheduler)
+            saved_checkpoint = True
+
+        # Exiting based on duration
+        if args.exit_duration_in_mins:
+            train_time = (time.time() - _TRAIN_START_TIME) / 60.0
+            done_cuda = torch.cuda.IntTensor(
+                [train_time > args.exit_duration_in_mins])
+            torch.distributed.all_reduce(
+                done_cuda, op=torch.distributed.ReduceOp.MAX)
+            done = done_cuda.item()
+            if done:
+                if not saved_checkpoint:
+                    save_checkpoint_and_time(iteration, model, optimizer,
+                                             lr_scheduler)
+                print_datetime('exiting program after {} minutes'.format(train_time))
+                sys.exit()
+
+        # Exiting based on iterations
+        if args.exit_interval and iteration % args.exit_interval == 0:
+            if not saved_checkpoint:
+                save_checkpoint_and_time(iteration, model, optimizer,
+                                         lr_scheduler)
+            torch.distributed.barrier()
+            print_datetime('exiting program at iteration {}'.format(iteration))
+            sys.exit()
+
+    return iteration
+
+
+def evaluate(forward_step_func, data_iterator, model, verbose=False):
+    """Evaluation."""
+    args = get_args()
+
+    # Turn on evaluation mode which disables dropout.
+    for model_module in model:
+        model_module.eval()
+
+    if args.curriculum_learning and \
+        args.pipeline_model_parallel_size >= 1:
+        # When curriculum learning is used with pipeline parallelism, we need
+        # this logic to ensure that the eval data is not truncated. If there
+        # is a seqlen change due to that, we need to call
+        # reset_activation_shape() to reset some buffers in deepspeed pipeline
+        # engine.
+        if args.curriculum_seqlen < args.seq_length:
+            args.curriculum_seqlen = args.seq_length
+            model[0].reset_activation_shape()
+
+    total_loss_dict = {}
+
+    with torch.no_grad():
+        iteration = 0
+        while iteration < args.eval_iters:
+            iteration += 1
+            if verbose and iteration % args.log_interval == 0:
+                print_rank_0('Evaluating iter {}/{}'.format(iteration,
+                                                            args.eval_iters))
+
+            if mpu.get_pipeline_model_parallel_world_size() > 1:
+                if args.virtual_pipeline_model_parallel_size is not None:
+                    forward_backward_func = forward_backward_pipelining_with_interleaving
+                else:
+                    forward_backward_func = forward_backward_pipelining_without_interleaving
+            else:
+                forward_backward_func = forward_backward_no_pipelining
+
+            if args.deepspeed:
+                # DeepSpeed uses eval_batch() and already aggregates losses.
+                assert isinstance(model, list) and len(model) == 1
+                loss = model[0].eval_batch(data_iterator)
+                loss_dicts = [{'lm loss' : loss}] * get_num_microbatches()
+            else:
+                loss_dicts = forward_backward_func(
+                    forward_step_func, data_iterator, model, optimizer=None,
+                    timers=None, forward_only=True)
+
+            if mpu.is_pipeline_last_stage(ignore_virtual=True):
+                # Reduce across processes.
+                for loss_dict in loss_dicts:
+                    for key in loss_dict:
+                        total_loss_dict[key] = total_loss_dict.get(
+                            key, torch.cuda.FloatTensor([0.0])) + loss_dict[key]
+
+            args.consumed_valid_samples += mpu.get_data_parallel_world_size() \
+                                           * args.micro_batch_size \
+                                           * get_num_microbatches()
+    # Move model back to the train mode.
+    for model_module in model:
+        model_module.train()
+
+    for key in total_loss_dict:
+        total_loss_dict[key] /= args.eval_iters * get_num_microbatches()
+
+    if args.curriculum_learning and \
+        args.pipeline_model_parallel_size >= 1:
+        # roll back to actual curriculum seqlen at the end of eval.
+        args.curriculum_seqlen = args.curriculum_scheduler.update_difficulty( \
+            args.iteration + 1)
+        if args.curriculum_seqlen < args.seq_length:
+            model[0].reset_activation_shape()
+
+    return total_loss_dict
+
+def evaluate_and_print_results(prefix, forward_step_func,
+                               data_iterator, model,
+                               iteration, verbose=False, **kwargs):
+    """Helper function to evaluate and dump results on screen."""
+
+
+    args = get_args()
+    writer = get_tensorboard_writer()
+
+    ds_name = kwargs.get("data_group_name", None)
+    # print corresponding dataset name (used for multiple validation datasets)
+    tf_plot_prefix = f"lm-loss-validation/{ds_name}" if ds_name else "lm-loss-validation"
+
+    total_loss_dict = evaluate(forward_step_func, data_iterator, model, verbose)
+    string = '{} loss at {} | '.format(ds_name, prefix) if ds_name is not None\
+        else 'validation loss at {} | '.format(prefix)
+    for key in total_loss_dict:
+        string += '{} value: {:.6E} | '.format(key, total_loss_dict[key].item())
+        ppl = math.exp(min(20, total_loss_dict[key].item()))
+        string += '{} PPL: {:.6E} | '.format(key, ppl)
+        if writer and is_last_rank():
+            writer.add_scalar(f'{tf_plot_prefix}/{key} validation',
+                              total_loss_dict[key].item(),
+                              iteration)
+            writer.add_scalar(f'{tf_plot_prefix}/{key} validation vs samples',
+                              total_loss_dict[key].item(),
+                              args.consumed_train_samples)
+            writer.add_scalar(f'{tf_plot_prefix}/{key} validation vs tokens',
+                              total_loss_dict[key].item(),
+                              args.consumed_train_tokens)
+            writer.add_scalar(f'{tf_plot_prefix}/{key} validation vs gigaflos (without embeddings)',
+                              total_loss_dict[key].item(),
+                              args.gigaflos_no_embeds)
+            if args.log_validation_ppl_to_tensorboard:
+                writer.add_scalar(f'{tf_plot_prefix}/{key} validation ppl', ppl,
+                                  iteration)
+                writer.add_scalar(f'{tf_plot_prefix}/{key} validation ppl vs samples',
+                                  ppl, args.consumed_train_samples)
+                writer.add_scalar(f'{tf_plot_prefix}/{key} validation ppl vs tokens',
+                                  ppl, args.consumed_train_tokens)
+                writer.add_scalar(f'{tf_plot_prefix}/{key} validation ppl vs gigaflos (without embeddings)',
+                                  ppl, args.gigaflos_no_embeds)
+
+    length = len(string) + 1
+    print_rank_last('-' * length)
+    print_rank_last(string)
+    print_rank_last('-' * length)
+
+
+def cyclic_iter(iter):
+    while True:
+        for x in iter:
+            yield x
+
+def build_train_valid_test_data_iterators(
+        build_train_valid_test_datasets_provider):
+    """XXX"""
+    args = get_args()
+
+    (train_dataloader, valid_dataloaders, test_dataloaders) = (None, None, None)
+
+    print_rank_0('> building train, validation, and test datasets ...')
+
+    # Backward compatibility, assume fixed batch size.
+    if args.iteration > 0 and args.consumed_train_samples == 0:
+        assert args.train_samples is None, \
+            'only backward compatiblity support for iteration-based training'
+        args.consumed_train_samples = args.iteration * args.global_batch_size
+    # it's possible that train was run, but not eval and it's valid if
+    # args.consumed_valid_samples == 0
+    # TODO: eval_interval could have changed between runs, so this might still be wrong
+    if args.iteration // args.eval_interval > 0 and args.consumed_valid_samples == 0:
+        assert args.train_samples is None, \
+            'only backward compatiblity support for iteration-based training'
+        args.consumed_valid_samples = (args.iteration // args.eval_interval) * \
+            args.eval_iters * args.global_batch_size
+
+    # Data loader only on rank 0 of each model parallel group.
+    if mpu.get_tensor_model_parallel_rank() == 0:
+        # Number of train/valid/test samples.
+        if args.train_samples:
+            train_samples = args.train_samples
+        else:
+            train_samples = args.train_iters * args.global_batch_size
+        eval_iters = (args.train_iters // args.eval_interval + 1) * \
+                     args.eval_iters
+        test_iters = args.eval_iters
+        train_val_test_num_samples = [train_samples,
+                                      eval_iters * args.global_batch_size,
+                                      test_iters * args.global_batch_size]
+        print_rank_0(' > datasets target sizes (minimum size):')
+        print_rank_0('    train:      {}'.format(train_val_test_num_samples[0]))
+        print_rank_0('    validation: {}'.format(train_val_test_num_samples[1]))
+        print_rank_0('    test:       {}'.format(train_val_test_num_samples[2]))
+
+        # Build the datasets.
+        train_ds, valid_ds, test_ds = build_train_valid_test_datasets_provider(train_val_test_num_samples)
+
+        # if dataloading option is not 2 convert to list to allow
+        # same interface for multiple data groups
+        # for validation and testing in option 2
+        if type(train_ds) != list and train_ds is not None:
+            train_ds = [train_ds]
+        if type(valid_ds) != list and valid_ds is not None:
+            valid_ds = [valid_ds]
+        if type(test_ds) != list and test_ds is not None:
+            test_ds = [test_ds]
+
+        # Build dataloders.
+        assert len(train_ds) == 1, "only one training dataset group is allowed"
+
+        # train_dataloader is a single item while valid_dataloaders
+        # and test_dataloaders are arrays
+        train_dataloader = build_pretraining_data_loader(
+            train_ds[0], args.consumed_train_samples)
+
+        # We collapse None and empty list as both should mean we don't run validation
+        # args.consumed_valid_samples accumulates the sum of valid steps for every dataset, which are all equal
+        #
+        # XXX: we get a deadlock in the dataloader on multi-dataset eval, after the first dataset,
+        # possibly due to this bug in pytorch https://github.com/pytorch/pytorch/pull/25158. Using
+        # num_workers=0 to work around it - the training can't use that since it impacts throughput
+        # by a few percent
+        valid_dataloaders = [build_pretraining_data_loader(d, args.consumed_valid_samples // len(valid_ds), num_workers=args.valid_num_workers)
+                            for d in valid_ds] \
+                            if valid_ds is not None else []
+        # We collapse None and empty list as both should mean we don't run test
+        test_dataloaders = [build_pretraining_data_loader(d, 0) for d in test_ds] \
+                            if test_ds is not None else []
+
+        # Flags to know if we need to do training/validation/testing.
+        do_train = train_dataloader is not None and args.train_iters > 0 and not args.eval_only
+
+        # Need to broadcast num_tokens and num_type_tokens.
+        flags = torch.cuda.LongTensor([
+            int(do_train),
+            len(valid_dataloaders) if args.eval_iters > 0 else 0, # eval_iters == 0 is equivalent to having no validation
+            len(test_dataloaders) if args.eval_iters > 0 else 0, # eval_iters == 0 is equivalent to having no test
+        ])
+    else:
+        flags = torch.cuda.LongTensor([0, 0, 0])
+
+    # Broadcast num tokens.
+    torch.distributed.broadcast(flags,
+                                mpu.get_tensor_model_parallel_src_rank(),
+                                group=mpu.get_tensor_model_parallel_group())
+    args.do_train = flags[0].item()
+    num_valid_ds = flags[1].item()
+    num_test_ds = flags[2].item()
+    assert num_test_ds >= 0
+    assert num_valid_ds >= 0
+    args.do_valid = num_valid_ds > 0
+    args.do_test = num_test_ds > 0
+
+    # Build iterators.
+    dl_type = args.dataloader_type
+    assert dl_type in ['single', 'cyclic']
+
+    if train_dataloader is not None:
+        train_data_iterator = iter(train_dataloader) if dl_type in ['single'] \
+                              else iter(cyclic_iter(train_dataloader))
+    else:
+        train_data_iterator = None
+
+    if valid_dataloaders is not None:
+        valid_data_iterators = [iter(vdl) if dl_type in ['single'] \
+                              else iter(cyclic_iter(valid_dataloaders))
+                                 for vdl in valid_dataloaders]
+    else:
+        valid_data_iterators = [None] * num_valid_ds
+
+    if test_dataloaders is not None:
+        test_data_iterators = [iter(tdl) if dl_type in ['single'] \
+                             else iter(cyclic_iter(test_dataloaders))
+                            for tdl in test_dataloaders]
+    else:
+        test_data_iterators = [None] * num_test_ds
+
+    return train_data_iterator, valid_data_iterators, test_data_iterators

megatron/utils.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""General utilities."""
+
+import os
+import sys
+import warnings
+from random import randint
+
+import torch
+from torch import nn
+from torch.nn.parallel import DistributedDataParallel as torchDDP
+
+from apex.multi_tensor_apply import multi_tensor_applier
+import amp_C
+
+from megatron import get_args, logging
+from megatron import print_rank_0
+from megatron import get_adlr_autoresume
+from megatron import mpu
+from megatron.model.module import param_is_not_shared
+from megatron.model.utils import log_debug_usage
+from megatron.mpu.layers import param_is_not_tensor_parallel_duplicate, VocabParallelEmbedding
+from megatron import get_num_microbatches
+
+logger = logging.get_logger(__name__)
+
+def unwrap_model(model, module_instances=(torchDDP)):
+    return_list = True
+    if not isinstance(model, list):
+        model = [model]
+        return_list = False
+    unwrapped_model = []
+    for model_module in model:
+        while isinstance(model_module, module_instances):
+            model_module = model_module.module
+        unwrapped_model.append(model_module)
+    if not return_list:
+        return unwrapped_model[0]
+    return unwrapped_model
+
+
+def calc_params_l2_norm(model):
+    """Calculate l2 norm of parameters """
+    args = get_args()
+    if not isinstance(model, list):
+        model = [model]
+    # Remove duplicate params.
+    params_data = []
+    for model_ in model:
+        for param in model_.parameters():
+            is_not_shared = param_is_not_shared(param)
+            is_not_tp_duplicate = param_is_not_tensor_parallel_duplicate(param)
+            if is_not_shared and is_not_tp_duplicate:
+                if args.bf16:
+                    params_data.append(param.data.float())
+                else:
+                    params_data.append(param.data)
+    # Calculate norm
+    dummy_overflow_buf = torch.cuda.IntTensor([0])
+    norm, _ = multi_tensor_applier(
+        amp_C.multi_tensor_l2norm,
+        dummy_overflow_buf,
+        [params_data],
+        False # no per-parameter norm
+    )
+    norm_2 = norm * norm
+    # Sum across all model-parallel GPUs.
+    torch.distributed.all_reduce(norm_2,
+                                 op=torch.distributed.ReduceOp.SUM,
+                                 group=mpu.get_model_parallel_group())
+    return norm_2.item() ** 0.5
+
+
+def average_losses_across_data_parallel_group(losses):
+    """Reduce a tensor of losses across all GPUs."""
+    averaged_losses = torch.cat(
+        [loss.clone().detach().view(1) for loss in losses])
+    torch.distributed.all_reduce(averaged_losses,
+                                 group=mpu.get_data_parallel_group())
+    averaged_losses = averaged_losses / \
+        torch.distributed.get_world_size(group=mpu.get_data_parallel_group())
+
+    return averaged_losses
+
+
+def report_memory(name):
+    """Simple GPU memory report."""
+    mega_bytes = 1024.0 * 1024.0
+    string = name + ' memory (MB)'
+    string += ' | allocated: {}'.format(
+        torch.cuda.memory_allocated() / mega_bytes)
+    string += ' | max allocated: {}'.format(
+        torch.cuda.max_memory_allocated() / mega_bytes)
+    string += ' | reserved: {}'.format(
+        torch.cuda.memory_reserved() / mega_bytes)
+    string += ' | max reserved: {}'.format(
+        torch.cuda.max_memory_reserved() / mega_bytes)
+    if mpu.get_data_parallel_rank() == 0:
+        print("[Rank {}] {}".format(torch.distributed.get_rank(), string),
+              flush=True)
+
+
+def print_params_min_max_norm(optimizer, iteration):
+    """Print min, max, and norm of all parameters."""
+    index = 0
+    rank = torch.distributed.get_rank()
+    string = 'iteration, rank, index, tensor-model-parallel, min, max, norm\n'
+    optimizer_ = optimizer.optimizer
+    for param_group in optimizer_.param_groups:
+        for param in param_group['params']:
+            index += 1
+            min_ = param.data.min()
+            max_ = param.data.max()
+            norm = torch.linalg.norm(param.data)
+            string += '{:7d}, {:4d}, {:4d}, {:2d}, '.format(
+                iteration, rank, index, int(param.tensor_model_parallel))
+            string += '{:.6E}, {:.6E}, {:.6E}\n'.format(min_, max_, norm)
+    print(string, flush=True)
+
+
+def check_adlr_autoresume_termination(iteration, model,
+                                      optimizer, lr_scheduler):
+    """Check for autoresume signal and exit if it is received."""
+    from megatron.checkpointing import save_checkpoint
+
+    args = get_args()
+    autoresume = get_adlr_autoresume()
+    # Add barrier to ensure consistnecy.
+    torch.distributed.barrier()
+    if autoresume.termination_requested():
+        if args.save:
+            save_checkpoint(iteration, model, optimizer, lr_scheduler)
+        print_rank_0(">>> autoresume termination request found!")
+        if torch.distributed.get_rank() == 0:
+            autoresume.request_resume()
+        print_rank_0(">>> training terminated. Returning")
+        sys.exit(0)
+
+
+def get_ltor_masks_and_position_ids(
+        data,
+        eod_token,
+        reset_position_ids,
+        reset_attention_mask,
+        eod_mask_loss,
+        prefix_indices,
+        loss_on_targets_only,
+    ):
+    """
+    Build masks and position id for left to right model.
+    :param prefix_indices: argument can have multiple types:
+        - None signifies that the model is fully autoregressive.
+        - List[int] the argument holds all prefix indices that split a row into an input and a target
+        - List[List[int]] the argument holds all prefix indices that split documents between input and target.
+    :param loss_on_targets_only: bool to determine if we should mask loss on prefix.
+    """
+
+    # Extract batch size and sequence length.
+    micro_batch_size, seq_length = data.size()
+
+    # Attention mask (lower triangular).
+    if reset_attention_mask or prefix_indices is not None:
+        att_mask_batch = micro_batch_size
+    else:
+        att_mask_batch = 1
+    attention_mask = torch.tril(torch.ones(
+        (att_mask_batch, seq_length, seq_length), device=data.device)).view(
+            att_mask_batch, 1, seq_length, seq_length)
+
+    # Loss mask.
+    loss_mask = torch.ones(data.size(), dtype=torch.float, device=data.device)
+    if eod_mask_loss:
+        loss_mask[data == eod_token] = 0.0
+
+    # Position ids.
+    position_ids = torch.arange(seq_length, dtype=torch.long,
+                                device=data.device)
+    position_ids = position_ids.unsqueeze(0).expand_as(data)
+    # We need to clone as the ids will be modifed based on batch index.
+    if reset_position_ids:
+        position_ids = position_ids.clone()
+
+    if reset_position_ids or reset_attention_mask or prefix_indices is not None:
+        # Loop through the batches:
+        for b in range(micro_batch_size):
+
+            # Find indecies where EOD token is.
+            eod_index = position_ids[b, data[b] == eod_token]
+
+            # If the last eod token is not the last token of the sequence, we suppose that there is a partial document
+            # We treat this case as if we add an eod token at the end of the sequence.
+            if data[b][-1] != eod_token:
+                eod_index = torch.cat(
+                    (eod_index, torch.tensor([len(data[b])], dtype=eod_index.dtype, device=eod_index.device))
+                )
+
+            # Detach indecies from positions if going to modify positions.
+            if reset_position_ids:
+                eod_index = eod_index.clone()
+
+            # Loop through EOD indecies:
+            prev_index = 0
+            for j in range(eod_index.size()[0]):
+                i = eod_index[j]
+
+                if reset_attention_mask:
+                    # Prevent cross document interactions.
+                    attention_mask[b, 0, (i + 1):, :(i + 1)] = 0
+
+                    # Prefix lm per document.
+                    if prefix_indices:
+                        assert isinstance(prefix_indices[b], list), f"prefix for a row has to be document specific, and consequently return a list, got {prefix_indices[b]}"
+                        attention_mask[b, 0, prev_index: prefix_indices[b][j], prev_index: prefix_indices[b][j]] = 1
+                        if loss_on_targets_only:
+                            # Last token of the prefix should predict the prefix_index id
+                            loss_mask[b, prev_index: prefix_indices[b][j] - 1] = 0.0
+
+                # Reset positions.
+                if reset_position_ids:
+                    position_ids[b, (i + 1):] -= (i + 1 - prev_index)
+
+                prev_index = i + 1
+
+            # Prefix lm per row.
+            if prefix_indices is not None and (reset_attention_mask is False):
+                assert isinstance(prefix_indices[b], int), \
+                    f"prefix for a row has to be row specific, and consequently return an int, got {prefix_indices[b]}"
+                attention_mask[b, 0, :prefix_indices[b], :prefix_indices[b]] = 1
+                if loss_on_targets_only:
+                    # Last token of the prefix should predict the prefix_index id
+                    loss_mask[b, :prefix_indices[b] - 1] = 0.0
+
+    # Convert attention mask to binary:
+    attention_mask = (attention_mask < 0.5)
+
+    return attention_mask, loss_mask, position_ids
+
+
+def get_packed_attention_mask(is_causal: bool, causal_mask: torch.Tensor, decoder_is_inputs: torch.Tensor, segment_ids: torch.Tensor):
+    """
+    Inspired by https://github.com/google-research/t5x/blob/7193407f98a8b18100b71a04ff777238be1682ca/t5x/examples/decoder_only/layers.py#L978
+
+    Arguments:
+        - is_causal: determines if the masking should be causal in the `inputs` part
+        - causal_mask: torch.BoolTensor [batch_size, sequence_length, sequence_length]
+        - decoder_is_inputs: torch.BoolTensor [batch_size, sequence_length]
+        - segment_ids: torch.IntTensor [batch_size, sequence_length]
+    Returns:
+        - attention_mask: torch.BoolTensor [batch_size, 1, sequence_length, sequence_length]
+    """
+
+    """Causal Inputs Mask:
+    mask = [[[[1, 1, 0, 0, 0, 0, 0],
+            [1, 1, 0, 0, 0, 0, 0],
+            [1, 1, 1, 0, 0, 0, 0],
+            [1, 1, 1, 1, 1, 0, 0],
+            [1, 1, 1, 1, 1, 0, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 1]]]]
+    """
+    assert causal_mask.dtype == torch.bool
+    assert segment_ids.dtype == torch.long
+    if is_causal:
+        causal_inputs_mask = causal_mask
+    else:
+        assert decoder_is_inputs.dtype == torch.bool
+        inputs_mask = decoder_is_inputs[:, None, :, None] * decoder_is_inputs[:, None, None, :]
+        causal_inputs_mask = causal_mask + inputs_mask
+
+    """Padding Mask:
+    mask = [[[[1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [1, 1, 1, 1, 1, 1, 0],
+            [0, 0, 0, 0, 0, 0, 0]]]]
+    """
+    padding_mask = (segment_ids != 0)[:, None, :, None] * (segment_ids != 0)[:, None, None, :]
+
+    """Segment Mask:
+    mask = [[[[1, 1, 1, 0, 0, 0, 0],
+            [1, 1, 1, 0, 0, 0, 0],
+            [1, 1, 1, 0, 0, 0, 0],
+            [0, 0, 0, 1, 1, 1, 0],
+            [0, 0, 0, 1, 1, 1, 0],
+            [0, 0, 0, 1, 1, 1, 0],
+            [0, 0, 0, 0, 0, 0, 0]]]]
+    """
+    segment_mask = segment_ids[:, None, :, None] == segment_ids[:, None, None, :]
+
+    """Final Mask:
+    mask = [[[[1, 1, 0, 0, 0, 0, 0],
+            [1, 1, 0, 0, 0, 0, 0],
+            [1, 1, 1, 0, 0, 0, 0],
+            [0, 0, 0, 1, 1, 0, 0],
+            [0, 0, 0, 1, 1, 0, 0],
+            [0, 0, 0, 1, 1, 1, 0],
+            [0, 0, 0, 0, 0, 0, 0]]]]
+    """
+    attention_mask = causal_inputs_mask * padding_mask * segment_mask
+
+    # Convert attention mask to binary:
+    attention_mask = (attention_mask < 0.5)
+
+    return attention_mask
+
+def param_size(parameter):
+    return parameter.ds_numel if hasattr(parameter, 'ds_id') else parameter.nelement()
+
+
+def unique_param_count(param_list):
+    # not actually deduplicating tied variables for now (which causes the PP > 1 double-counting bug)
+    return sum(dict((p.data_ptr(), param_size(p)) for p in param_list).values())
+
+
+def non_embedding_params(module):
+    embedding_param_names = [
+        f"{name}.weight" for name, module_type in module.named_modules() if isinstance(module_type, nn.Embedding) or isinstance(module_type, VocabParallelEmbedding)
+    ]
+    non_embedding_parameters = [
+        parameter for name, parameter in module.named_parameters() if name not in embedding_param_names
+    ]
+    return unique_param_count(non_embedding_parameters)
+
+
+def get_parameters_in_billions(model, exclude_embeddings=False):
+    gpus_per_model = torch.distributed.get_world_size(group=mpu.get_model_parallel_group())
+
+    if exclude_embeddings:
+        approx_parameters_in_billions = sum([non_embedding_params(model_module) for model_module in model])
+    else:
+        args = get_args()
+        if args.rank == 0:
+            warnings.warn("Parameter count with the embeddings will be inaccurate with PP > 1, as the first and last stage hold several copies of the embeddings")
+        approx_parameters_in_billions = unique_param_count([p for model_module in model for p in model_module.parameters()])
+
+    return approx_parameters_in_billions*gpus_per_model/(1e9)
+
+
+def flops_calculator(model, args, iteration_time):
+    return # currently broken
+    gpus_per_model = torch.distributed.get_world_size(group = mpu.get_model_parallel_group())
+
+    approx_parameters_in_billions = get_parameters_in_billions(model)
+
+    batch_size = args.micro_batch_size * get_num_microbatches()
+
+    giga_flops_per_model_per_train_step = approx_parameters_in_billions * batch_size * args.seq_length * 2.0 * 4.0
+
+    effective_tera_flops_per_gpu = giga_flops_per_model_per_train_step / (iteration_time * 1000.0 * gpus_per_model)
+
+    print_rank_0(f"Effective Tera Flops per GPU: {round(effective_tera_flops_per_gpu, 2)} and total parameters {round(approx_parameters_in_billions, 3)} B")
+
+def get_prefix_indices(data, eod_token, partial_prefix_indices, reset_attention_mask):
+    """
+    Helper function in order to:
+     - randomly choose prefix index when there's no constraint
+     - check that prefix are compatible with convention.
+
+    :param data: torch.Tensor
+    :param eod_token: int, token_id used to signal end of document
+    :param partial_prefix_indices: this agument can have multiple types:
+        - None, it signals that all prefix indices are randomly sampled.
+        - List[Optional[int]], its length has to be equal to mini batch size. It stores all the indices for per row prefix.
+            Optional means that if set to None, we allows ourselves to sample one randomly.
+        - List[List[Optional[int]]], it follows the following rules:
+            - The first dimension refers to that sample, ie len(partial_prefix_indices) == len(data)
+            - The second dimension refers to the number of document of that sample, ie
+                len(partial_prefix_indices[b]) == (data[b] == eod_token).sum() (+1 for the last partial document).
+            - partial_prefix_indices have to be interleaved with eod_indices, ie
+                eod_indices[b][d-1] < partial_prefix_indices[b][d] < eod_indices[b][d] + 1 or is None.
+            - Optional means that if set to None, we allows ourselves to sample one randomly.
+    :param reset_attention_mask: bool, determines if prefixes are to be per document or per row.
+    :return Depending if prefix is per document or per row, the method returns:
+        - List[List[int]]: prefix indices for each document in case of per document prefix
+        - List[int]: prefix indices for rows else.
+    """
+    micro_batch_size, seq_length = data.size()
+    prefix_indices = []
+
+    assert partial_prefix_indices is None or len(partial_prefix_indices) == micro_batch_size, f"partial_prefix_indices has to be None or its length equal to {micro_batch_size}, got {len(partial_prefix_indices)}"
+    for batch_id in range(micro_batch_size):
+        # Prefix lm per document.
+        if reset_attention_mask:
+            prefix_indices.append([])
+
+            # Compute the index of all eod tokens in data.
+            eod_indices = (data[batch_id] == eod_token).nonzero().squeeze(-1)
+
+            # If the last eod token is not the last token of the sequence, we suppose that there is a partial document
+            # We treat this case as if we add an eod token at the end of the sequence.
+            if data[batch_id][-1] != eod_token:
+                eod_indices = torch.cat(
+                    (eod_indices,
+                     torch.tensor([len(data[batch_id])], dtype=eod_indices.dtype, device=eod_indices.device))
+                )
+
+            prev_index = 0
+            assert partial_prefix_indices is None or len(partial_prefix_indices[batch_id]) == len(eod_indices), f"The number of prefixes has to match the number of documents, complete or partial. Got {len(partial_prefix_indices[batch_id])} prefixes and {len(eod_indices)} documents"
+
+            for doc_id, eod_index in enumerate(eod_indices):
+                assert partial_prefix_indices is None or isinstance(partial_prefix_indices[batch_id], list), f"Per document prefix has to store a list on indices for each row, got {partial_prefix_indices[batch_id]}"
+                # Prefix index is defined as the first index that isn't attended by all tokens in a document
+                if partial_prefix_indices is None or partial_prefix_indices[batch_id][doc_id] is None:
+                    # We need to randomly generate a prefix index that satisfies the interleave condition in the docstring
+                    prefix_index = randint(prev_index + 1, eod_index)
+                else:
+                    # We get value from partial_prefix_indices, and run validation on that value
+                    prefix_index = partial_prefix_indices[batch_id][doc_id]
+                assert prev_index + 1 <= prefix_index <= eod_index, f"Prefix index needs to be between documents indices, {prev_index + 1} <= {prefix_index} <= {eod_index} should be True."
+
+                prefix_indices[batch_id].append(prefix_index)
+                prev_index = eod_index + 1
+
+        # Prefix lm per row.
+        else:
+            assert partial_prefix_indices is None or isinstance(partial_prefix_indices[batch_id], int), \
+                f"Per document prefix has to store an int for each row, got {partial_prefix_indices[batch_id]}"
+
+            # Prefix index is defined as the first index that isn't attended by all previous tokens in a document
+            prefix_index: int
+            if partial_prefix_indices is None or partial_prefix_indices[batch_id] is None:
+                # 0 being the first prefix index makes no sense since 0 always attends to itself, and there are no other tokens before.
+                prefix_index = randint(1, seq_length)
+            else:
+                # We get value from partial_prefix_indices, and run validation on that value
+                prefix_index = partial_prefix_indices[batch_id]
+            assert 1 <= prefix_index <= seq_length, f"Prefix index needs to be between documents indices, 1 <= {prefix_index} <= {seq_length} should be True."
+            prefix_indices.append(prefix_index)
+
+    return prefix_indices
+
+
+@log_debug_usage(logger, "Using loss reweighting")
+def reweight_loss_mask_(loss_mask: torch.Tensor, tokens: torch.Tensor):
+    """Reweight loss mask in-place"""
+    _, seq_length = tokens.shape
+    weight_loss = torch.arange(seq_length, 0, -1, dtype=torch.float, device=loss_mask.device) / (seq_length + 1) * 2
+    # in-place operation
+    loss_mask *= weight_loss[None, :]
+
+
+def found_kill_switch():
+    args = get_args()
+    if args.kill_switch_path is not None and os.path.exists(args.kill_switch_path):
+        return True
+    else:
+        return False
+
+def get_fingerprint_header():
+    return f"{'min':^13} {'max':^13} {'mean':^13} {'l2 norm':^12} metadata"
+
+def get_fingerprint(p):
+    return f"{p.min():13.6e} {p.max():13.6e} {p.mean():13.6e} {p.norm():12.6e}"
+
+
+def dump_weights(preamble, iteration, model, optimizer, tensor=None):   
+    tp_rank = mpu.get_tensor_model_parallel_rank()
+    pp_rank = mpu.get_pipeline_model_parallel_rank()
+    dp_rank = mpu.get_data_parallel_rank()
+    dp_size = mpu.get_data_parallel_world_size()
+    fn = f"debug-bf16-{iteration}-pp{pp_rank}-tp{tp_rank}-dp{dp_rank}-{preamble}.txt"
+
+    # only care for first and last pp stages and dp0 tp0
+    #if not (mpu.is_pipeline_first_stage() or mpu.is_pipeline_last_stage()):
+    #    return
+
+    #if not (tp_rank == 0 and dp_rank == 0):
+    #    return
+
+    if tensor is not None:
+        orig_tensor = tensor
+        if hasattr(tensor, "_hp_param"):
+            numel = tensor._hp_param.numel() # // dp_size
+            tensor = tensor.flatten().narrow(0, 0, numel)
+
+    #print(fn)
+    with open(fn, "w") as fh:
+        fh.write(f"{get_fingerprint_header()}\n")
+
+        if tensor is not None:
+            fh.write(f"{get_fingerprint(tensor)} tensor {tensor.shape}\n")
+        else:
+            for n, p in model[0].named_parameters():
+                fh.write(f"{get_fingerprint(p)} {n} {p.shape}\n")
+
+
+    return 
+
+
+    # until we figure out how to dump the actual fp32 values don't do this
+    fn = f"debug-fp32-{iteration}-pp{pp_rank}-tp{tp_rank}-dp{dp_rank}-{preamble}.txt"
+    with open(fn, "w") as fh:
+        fh.write(f"{get_fingerprint_header()}\n")
+        if tensor is not None:
+            tensor = orig_tensor
+            if hasattr(tensor, "_hp_param"):
+                fh.write(f"{get_fingerprint(tensor._hp_param)} tensor {tensor._hp_param.shape}\n")
+                #fh.write(f"{get_fingerprint(tensor._hp_grad)} tensor grad\n")
+            else:
+                fh.write(f"{get_fingerprint(tensor)} tensor {tensor.shape}\n")
+                #fh.write(f"{get_fingerprint(tensor.grad)} tensor grad\n")
+
+        else:
+            if hasattr(model[0].module.tied_modules, "embed"):
+                p = model[0].module.tied_modules.embed.word_embeddings.weight._hp_param
+                fh.write(f"{get_fingerprint(p)} module.tied_modules.embed.word_embeddings.weight._hp_param {p.shape}\n")
+
+        # for i, param_group in enumerate(optimizer.param_groups):
+        #     fh.write(f"{get_fingerprint(optimizer.fp32_groups_flat_partition[i])} group={i}\n")
+            #fh.write(f"{i}={optimizer.fp32_groups_flat_partition[i]}\n")
+    #     if mpu.is_pipeline_first_stage():
+    #         x = optimizer.fp32_groups_flat_partition[0]
+    #         fh.write(f"fp32={x[:402432]}\n")
+    #     if mpu.is_pipeline_last_stage()):
+    #         x = optimizer.fp32_groups_flat_partition[1]
+    #         fh.write(f"fp32={x[-402432:]}\n")
+
+    # import os
+    # import socket
+    # hostname = socket.gethostname()
+    # pid = os.getpid()
+    # global_rank = torch.distributed.get_rank()
+    #fn = f"debug-{iteration}-pp{pp_rank}-tp{tp_rank}-dp{dp_rank}-global{global_rank}-{preamble}-{pid}.txt"
\ No newline at end of file

pretrain_bert.py

+# coding=utf-8
+# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Pretrain BERT"""
+
+from functools import partial
+
+import torch
+import torch.nn.functional as F
+
+from megatron import get_args
+from megatron import print_rank_0
+from megatron import get_timers
+from megatron import mpu
+from megatron.data.dataset_utils import build_train_valid_test_datasets
+from megatron.model.bert_model import BertModel
+from megatron.training import pretrain
+from megatron.utils import average_losses_across_data_parallel_group
+
+
+def model_provider(pre_process=True, post_process=True):
+    """Build the model."""
+
+    print_rank_0('building BERT model ...')
+
+    args = get_args()
+    num_tokentypes = 2 if args.bert_binary_head else 0
+    model = BertModel(
+        num_tokentypes=num_tokentypes,
+        add_binary_head=args.bert_binary_head,
+        parallel_output=True,
+        pre_process=pre_process,
+        post_process=post_process)
+
+    return model
+
+
+def get_batch(data_iterator):
+    """Build the batch."""
+
+    # Items and their type.
+    keys = ['text', 'types', 'labels', 'is_random', 'loss_mask', 'padding_mask']
+    datatype = torch.int64
+
+    # Broadcast data.
+    if data_iterator is not None:
+        data = next(data_iterator)
+    else:
+        data = None
+    data_b = mpu.broadcast_data(keys, data, datatype)
+
+    # Unpack.
+    tokens = data_b['text'].long()
+    types = data_b['types'].long()
+    sentence_order = data_b['is_random'].long()
+    loss_mask = data_b['loss_mask'].float()
+    lm_labels = data_b['labels'].long()
+    padding_mask = data_b['padding_mask'].long()
+
+    return tokens, types, sentence_order, loss_mask, lm_labels, padding_mask
+
+
+def loss_func(loss_mask, sentence_order, output_tensor):
+    lm_loss_, sop_logits = output_tensor
+
+    lm_loss_ = lm_loss_.float()
+    loss_mask = loss_mask.float()
+    lm_loss = torch.sum(
+        lm_loss_.view(-1) * loss_mask.reshape(-1)) / loss_mask.sum()
+
+    if sop_logits is not None:
+        sop_loss = F.cross_entropy(sop_logits.view(-1, 2).float(),
+                                   sentence_order.view(-1),
+                                   ignore_index=-1)
+        sop_loss = sop_loss.float()
+        loss = lm_loss + sop_loss
+        averaged_losses = average_losses_across_data_parallel_group(
+            [lm_loss, sop_loss])
+        return loss, {'lm loss': averaged_losses[0],
+                      'sop loss': averaged_losses[1]}
+
+    else:
+        loss = lm_loss
+        averaged_losses = average_losses_across_data_parallel_group(
+            [lm_loss])
+        return loss, {'lm loss': averaged_losses[0]}
+
+
+def forward_step(data_iterator, model):
+    """Forward step."""
+    args = get_args()
+    timers = get_timers()
+
+    # Get the batch.
+    timers('batch-generator').start()
+    tokens, types, sentence_order, loss_mask, lm_labels, padding_mask = get_batch(
+        data_iterator)
+    timers('batch-generator').stop()
+
+    if not args.bert_binary_head:
+        types = None
+
+    # Forward pass through the model.
+    output_tensor = model(tokens, padding_mask, tokentype_ids=types,
+                          lm_labels=lm_labels)
+
+    return output_tensor, partial(loss_func, loss_mask, sentence_order)
+
+
+def train_valid_test_datasets_provider(train_val_test_num_samples):
+    """Build train, valid, and test datasets."""
+    args = get_args()
+
+    print_rank_0('> building train, validation, and test datasets '
+                 'for BERT ...')
+    train_ds, valid_ds, test_ds = build_train_valid_test_datasets(
+        data_prefix=args.data_path,
+        data_impl=args.data_impl,
+        splits_string=args.split,
+        train_valid_test_num_samples=train_val_test_num_samples,
+        max_seq_length=args.seq_length,
+        masked_lm_prob=args.mask_prob,
+        short_seq_prob=args.short_seq_prob,
+        seed=args.seed,
+        skip_warmup=(not args.mmap_warmup),
+        binary_head=args.bert_binary_head)
+    print_rank_0("> finished creating BERT datasets ...")
+
+    return train_ds, valid_ds, test_ds
+
+
+if __name__ == "__main__":
+
+    pretrain(train_valid_test_datasets_provider, model_provider, forward_step,
+             args_defaults={'tokenizer_type': 'BertWordPieceLowerCase'})