MANIFEST.in README.md setup.py /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/colossal_C_frontend.cpp /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/layer_norm_hip.cpp /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/layer_norm_hip_kernel.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multi_tensor_adam.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multi_tensor_l2norm_kernel.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multi_tensor_lamb.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multi_tensor_scale_kernel.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multi_tensor_sgd_kernel.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/multihead_attention_1d.cpp /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/scaled_masked_softmax.cpp /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/scaled_masked_softmax_hip.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax.cpp /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax_hip.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/cublas_wrappers.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/dropout_kernels.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/general_kernels.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/hip_util.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/normalize_kernels.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/softmax_kernels.hip /public/home/huchen/colossalAI/ColossalAI/colossalai/kernel/hip_native/csrc/kernels/transform_kernels.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/colossal_C_frontend.cpp /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/layer_norm_hip.cpp /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/layer_norm_hip_kernel.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multi_tensor_adam.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multi_tensor_l2norm_kernel.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multi_tensor_lamb.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multi_tensor_scale_kernel.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multi_tensor_sgd_kernel.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/multihead_attention_1d.cpp /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/scaled_masked_softmax.cpp /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/scaled_masked_softmax_hip.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax.cpp /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax_hip.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/cublas_wrappers.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/dropout_kernels.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/general_kernels.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/hip_util.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/normalize_kernels.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/softmax_kernels.hip /public/home/huchen/colossalAI/colossalai_dtk2210/colossalai/kernel/hip_native/csrc/kernels/transform_kernels.hip colossalai/__init__.py colossalai/constants.py colossalai/core.py colossalai/global_variables.py colossalai/initialize.py colossalai.egg-info/PKG-INFO colossalai.egg-info/SOURCES.txt colossalai.egg-info/dependency_links.txt colossalai.egg-info/requires.txt colossalai.egg-info/top_level.txt colossalai/amp/__init__.py colossalai/amp/amp_type.py colossalai/amp/apex_amp/__init__.py colossalai/amp/apex_amp/apex_amp.py colossalai/amp/naive_amp/__init__.py colossalai/amp/naive_amp/_fp16_optimizer.py colossalai/amp/naive_amp/naive_amp.py colossalai/amp/torch_amp/__init__.py colossalai/amp/torch_amp/_grad_scaler.py colossalai/amp/torch_amp/torch_amp.py colossalai/builder/__init__.py colossalai/builder/builder.py colossalai/builder/pipeline.py colossalai/communication/__init__.py colossalai/communication/collective.py colossalai/communication/p2p.py colossalai/communication/ring.py colossalai/communication/utils.py colossalai/context/__init__.py colossalai/context/config.py colossalai/context/parallel_context.py colossalai/context/parallel_mode.py colossalai/context/process_group_initializer/__init__.py colossalai/context/process_group_initializer/initializer_1d.py colossalai/context/process_group_initializer/initializer_2d.py colossalai/context/process_group_initializer/initializer_2p5d.py colossalai/context/process_group_initializer/initializer_3d.py colossalai/context/process_group_initializer/initializer_data.py colossalai/context/process_group_initializer/initializer_model.py colossalai/context/process_group_initializer/initializer_moe.py colossalai/context/process_group_initializer/initializer_pipeline.py colossalai/context/process_group_initializer/initializer_sequence.py colossalai/context/process_group_initializer/initializer_tensor.py colossalai/context/process_group_initializer/process_group_initializer.py colossalai/context/random/__init__.py colossalai/context/random/_helper.py colossalai/context/random/seed_manager.py colossalai/engine/__init__.py colossalai/engine/_base_engine.py colossalai/engine/gradient_handler/__init__.py colossalai/engine/gradient_handler/_base_gradient_handler.py colossalai/engine/gradient_handler/_data_parallel_gradient_handler.py colossalai/engine/gradient_handler/_moe_gradient_handler.py colossalai/engine/gradient_handler/_pipeline_parallel_gradient_handler.py colossalai/engine/gradient_handler/_sequence_parallel_gradient_handler.py colossalai/engine/gradient_handler/_zero_gradient_handler.py colossalai/engine/ophooks/__init__.py colossalai/engine/ophooks/_base_ophook.py colossalai/engine/ophooks/_memtracer_ophook.py colossalai/engine/schedule/__init__.py colossalai/engine/schedule/_base_schedule.py colossalai/engine/schedule/_non_pipeline_schedule.py colossalai/engine/schedule/_pipeline_schedule.py colossalai/kernel/__init__.py colossalai/kernel/cuda_native/__init__.py colossalai/kernel/cuda_native/layer_norm.py colossalai/kernel/cuda_native/multihead_attention.py colossalai/kernel/cuda_native/scaled_softmax.py colossalai/kernel/cuda_native/csrc/colossal_C_frontend.cpp colossalai/kernel/cuda_native/csrc/compat.h colossalai/kernel/cuda_native/csrc/layer_norm_cuda.cpp colossalai/kernel/cuda_native/csrc/layer_norm_cuda_kernel.cu colossalai/kernel/cuda_native/csrc/multi_tensor_adam.cu colossalai/kernel/cuda_native/csrc/multi_tensor_apply.cuh colossalai/kernel/cuda_native/csrc/multi_tensor_l2norm_kernel.cu colossalai/kernel/cuda_native/csrc/multi_tensor_lamb.cu colossalai/kernel/cuda_native/csrc/multi_tensor_scale_kernel.cu colossalai/kernel/cuda_native/csrc/multi_tensor_sgd_kernel.cu colossalai/kernel/cuda_native/csrc/multihead_attention_1d.cpp colossalai/kernel/cuda_native/csrc/multihead_attention_1d.h colossalai/kernel/cuda_native/csrc/scaled_masked_softmax.cpp colossalai/kernel/cuda_native/csrc/scaled_masked_softmax.h colossalai/kernel/cuda_native/csrc/scaled_masked_softmax_cuda.cu colossalai/kernel/cuda_native/csrc/scaled_upper_triang_masked_softmax.cpp colossalai/kernel/cuda_native/csrc/scaled_upper_triang_masked_softmax.h colossalai/kernel/cuda_native/csrc/scaled_upper_triang_masked_softmax_cuda.cu colossalai/kernel/cuda_native/csrc/type_shim.h colossalai/kernel/cuda_native/csrc/kernels/cross_entropy.cu colossalai/kernel/cuda_native/csrc/kernels/cublas_wrappers.cu colossalai/kernel/cuda_native/csrc/kernels/cuda_util.cu colossalai/kernel/cuda_native/csrc/kernels/dropout_kernels.cu colossalai/kernel/cuda_native/csrc/kernels/general_kernels.cu colossalai/kernel/cuda_native/csrc/kernels/normalize_kernels.cu colossalai/kernel/cuda_native/csrc/kernels/softmax_kernels.cu colossalai/kernel/cuda_native/csrc/kernels/transform_kernels.cu colossalai/kernel/cuda_native/csrc/kernels/include/block_reduce.h colossalai/kernel/cuda_native/csrc/kernels/include/context.h colossalai/kernel/cuda_native/csrc/kernels/include/cross_entropy_layer.h colossalai/kernel/cuda_native/csrc/kernels/include/cublas_wrappers.h colossalai/kernel/cuda_native/csrc/kernels/include/cuda_util.h colossalai/kernel/cuda_native/csrc/kernels/include/dropout.h colossalai/kernel/cuda_native/csrc/kernels/include/feed_forward.h colossalai/kernel/cuda_native/csrc/kernels/include/kernels.h colossalai/kernel/cuda_native/csrc/kernels/include/ls_cub.cuh colossalai/kernel/cuda_native/csrc/kernels/include/normalize_layer.h colossalai/kernel/cuda_native/csrc/kernels/include/softmax.h colossalai/kernel/cuda_native/csrc/kernels/include/strided_batch_gemm.h colossalai/kernel/hip_native/csrc/colossal_C_frontend.cpp colossalai/kernel/hip_native/csrc/compat.h colossalai/kernel/hip_native/csrc/layer_norm_hip.cpp colossalai/kernel/hip_native/csrc/multi_tensor_apply.cuh colossalai/kernel/hip_native/csrc/multihead_attention_1d.cpp colossalai/kernel/hip_native/csrc/multihead_attention_1d.h colossalai/kernel/hip_native/csrc/scaled_masked_softmax.cpp colossalai/kernel/hip_native/csrc/scaled_masked_softmax.h colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax.cpp colossalai/kernel/hip_native/csrc/scaled_upper_triang_masked_softmax.h colossalai/kernel/hip_native/csrc/type_shim.h colossalai/kernel/hip_native/csrc/kernels/include/block_reduce.h colossalai/kernel/hip_native/csrc/kernels/include/context.h colossalai/kernel/hip_native/csrc/kernels/include/cross_entropy_layer.h colossalai/kernel/hip_native/csrc/kernels/include/cublas_wrappers.h colossalai/kernel/hip_native/csrc/kernels/include/dropout.h colossalai/kernel/hip_native/csrc/kernels/include/feed_forward.h colossalai/kernel/hip_native/csrc/kernels/include/hip_util.h colossalai/kernel/hip_native/csrc/kernels/include/kernels.h colossalai/kernel/hip_native/csrc/kernels/include/ls_cub.cuh colossalai/kernel/hip_native/csrc/kernels/include/normalize_layer.h colossalai/kernel/hip_native/csrc/kernels/include/softmax.h colossalai/kernel/hip_native/csrc/kernels/include/strided_batch_gemm.h colossalai/kernel/jit/__init__.py colossalai/kernel/jit/bias_dropout_add.py colossalai/kernel/jit/bias_gelu.py colossalai/kernel/jit/option.py colossalai/logging/__init__.py colossalai/logging/logging.py colossalai/nn/__init__.py colossalai/nn/init.py colossalai/nn/layer/__init__.py colossalai/nn/layer/base_layer.py colossalai/nn/layer/colossalai_layer/__init__.py colossalai/nn/layer/colossalai_layer/_utils.py colossalai/nn/layer/colossalai_layer/dropout.py colossalai/nn/layer/colossalai_layer/embedding.py colossalai/nn/layer/colossalai_layer/linear.py colossalai/nn/layer/colossalai_layer/normalization.py colossalai/nn/layer/moe/__init__.py colossalai/nn/layer/moe/_operation.py colossalai/nn/layer/moe/layers.py colossalai/nn/layer/parallel_1d/__init__.py colossalai/nn/layer/parallel_1d/_operation.py colossalai/nn/layer/parallel_1d/_utils.py colossalai/nn/layer/parallel_1d/layers.py colossalai/nn/layer/parallel_2d/__init__.py colossalai/nn/layer/parallel_2d/_operation.py colossalai/nn/layer/parallel_2d/_utils.py colossalai/nn/layer/parallel_2d/layers.py colossalai/nn/layer/parallel_2p5d/__init__.py colossalai/nn/layer/parallel_2p5d/_operation.py colossalai/nn/layer/parallel_2p5d/_utils.py colossalai/nn/layer/parallel_2p5d/layers.py colossalai/nn/layer/parallel_3d/__init__.py colossalai/nn/layer/parallel_3d/_operation.py colossalai/nn/layer/parallel_3d/_utils.py colossalai/nn/layer/parallel_3d/layers.py colossalai/nn/layer/parallel_sequence/__init__.py colossalai/nn/layer/parallel_sequence/_operation.py colossalai/nn/layer/parallel_sequence/_utils.py colossalai/nn/layer/parallel_sequence/layers.py colossalai/nn/layer/utils/__init__.py colossalai/nn/layer/utils/common.py colossalai/nn/layer/vanilla/__init__.py colossalai/nn/layer/vanilla/layers.py colossalai/nn/layer/wrapper/__init__.py colossalai/nn/layer/wrapper/lambda_wrapper.py colossalai/nn/layer/wrapper/pipeline_wrapper.py colossalai/nn/loss/__init__.py colossalai/nn/loss/loss_1d.py colossalai/nn/loss/loss_2d.py colossalai/nn/loss/loss_2p5d.py colossalai/nn/loss/loss_3d.py colossalai/nn/loss/loss_moe.py colossalai/nn/lr_scheduler/__init__.py colossalai/nn/lr_scheduler/cosine.py colossalai/nn/lr_scheduler/delayed.py colossalai/nn/lr_scheduler/linear.py colossalai/nn/lr_scheduler/multistep.py colossalai/nn/lr_scheduler/onecycle.py colossalai/nn/lr_scheduler/poly.py colossalai/nn/lr_scheduler/torch.py colossalai/nn/metric/__init__.py colossalai/nn/metric/_utils.py colossalai/nn/metric/accuracy_2d.py colossalai/nn/metric/accuracy_2p5d.py colossalai/nn/metric/accuracy_3d.py colossalai/nn/model/__init__.py colossalai/nn/model/model_from_config.py colossalai/nn/optimizer/__init__.py colossalai/nn/optimizer/colossalai_optimizer.py colossalai/nn/optimizer/fused_adam.py colossalai/nn/optimizer/fused_lamb.py colossalai/nn/optimizer/fused_sgd.py colossalai/nn/optimizer/lamb.py colossalai/nn/optimizer/lars.py colossalai/registry/__init__.py colossalai/registry/registry.py colossalai/trainer/__init__.py colossalai/trainer/_trainer.py colossalai/trainer/hooks/__init__.py colossalai/trainer/hooks/_base_hook.py colossalai/trainer/hooks/_checkpoint_hook.py colossalai/trainer/hooks/_log_hook.py colossalai/trainer/hooks/_lr_scheduler_hook.py colossalai/trainer/hooks/_metric_hook.py colossalai/utils/__init__.py colossalai/utils/activation_checkpoint.py colossalai/utils/checkpointing.py colossalai/utils/common.py colossalai/utils/cuda.py colossalai/utils/memory.py colossalai/utils/timer.py colossalai/utils/data_sampler/__init__.py colossalai/utils/data_sampler/base_sampler.py colossalai/utils/data_sampler/data_parallel_sampler.py colossalai/utils/gradient_accumulation/__init__.py colossalai/utils/gradient_accumulation/_gradient_accumulation.py colossalai/utils/multi_tensor_apply/__init__.py colossalai/utils/multi_tensor_apply/multi_tensor_apply.py colossalai/zero/__init__.py colossalai/zero/loss_scaler.py colossalai/zero/zero_redundancy_optimizer_level_2.py colossalai/zero/zero_redundancy_optimizer_level_3.py model_zoo/__init__.py model_zoo/helper.py model_zoo/bert/__init__.py model_zoo/gpt/__init__.py model_zoo/gpt/gpt.py model_zoo/mlp_mixer/__init__.py model_zoo/mlp_mixer/parallel_3d/__init__.py model_zoo/mlp_mixer/parallel_3d/mlp_mixer.py model_zoo/moe/__init__.py model_zoo/moe/models.py model_zoo/moe/util.py model_zoo/vit/__init__.py model_zoo/vit/vision_transformer_from_config.py model_zoo/vit/vit.py requirements/requirements-test.txt requirements/requirements-zero.txt requirements/requirements.txt