[gemini] add fake_release_chunk for keep-gathered chunk in the inference mode (#2671)

colossalai/gemini/chunk/manager.py

@@ -140,6 +140,14 @@ class ChunkManager:
         self.__add_memory_usage(chunk.memory_usage)
         return True
 
+    def fake_release_chunk(self, chunk: Chunk) -> None:
+        """Release gathered chunk in a fake mode.
+        This function is used for keep-gathered chunk in the inference mode.
+        """
+        assert chunk.keep_gathered
+        assert chunk.tensor_state_cnter[TensorState.HOLD] == chunk.num_tensors
+        self.__sub_accessed_chunk(chunk)
+
     def copy_tensor_to_chunk_slice(self, tensor: torch.Tensor, data: torch.Tensor) -> None:
         """
         Copy data to the chunk.

colossalai/nn/parallel/data_parallel.py

@@ -257,6 +257,9 @@ class ZeroDDP(ColoDDP):
         access_list = list(self.chunk_manager.accessed_chunks)
         # we need to scatter all accessed chunks and move them to their original places
         for chunk in access_list:
+            if chunk.keep_gathered:
+                self.chunk_manager.fake_release_chunk(chunk)
+            else:
                 assert chunk.can_release
                 self.chunk_manager.release_chunk(chunk)
             first_param = next(iter(chunk.tensors_info))

tests/test_gemini/update/test_inference.py

 from functools import partial
+from typing import Callable
 
 import pytest
 import torch
@@ -13,7 +14,7 @@ from colossalai.gemini.chunk import ChunkManager, init_chunk_manager, search_chu
 from colossalai.gemini.gemini_mgr import GeminiManager
 from colossalai.nn.optimizer import HybridAdam
 from colossalai.nn.optimizer.zero_optimizer import ZeroOptimizer
-from colossalai.nn.parallel import ZeroDDP
+from colossalai.nn.parallel import ZeroDDP, zero_model_wrapper
 from colossalai.testing import parameterize, rerun_if_address_is_in_use
 from colossalai.utils import free_port
 from colossalai.utils.cuda import get_current_device
@@ -36,9 +37,35 @@ def check_param(model: ZeroDDP, torch_model: torch.nn.Module):
         assert_close(value, temp_zero_value, rtol=1e-3, atol=4e-3)
 
 
+def multi_chunk_init(model: torch.nn.Module, placement_policy: str):
+    world_size = dist.get_world_size()
+    config_dict, *_ = search_chunk_configuration(model, search_range_mb=1, search_interval_byte=100)
+    config_dict[world_size]['chunk_size'] = 5000
+    config_dict[world_size]['keep_gathered'] = False
+    if placement_policy != 'cuda':
+        init_device = torch.device('cpu')
+    else:
+        init_device = None
+    chunk_manager = ChunkManager(config_dict, init_device=init_device)
+    gemini_manager = GeminiManager(placement_policy, chunk_manager)
+    model = ZeroDDP(model, gemini_manager, pin_memory=True)
+    return model
+
+
+def single_chunk_init(model: torch.nn.Module, placement_policy: str):
+    gemini_config = dict(
+        device=get_current_device(),
+        placement_policy=placement_policy,
+        pin_memory=True,
+    )
+    model = zero_model_wrapper(model=model, zero_stage=3, gemini_config=gemini_config)
+    return model
+
+
 @parameterize('placement_policy', ['cuda', 'cpu', 'auto', 'const'])
 @parameterize('model_name', ['gpt2'])
-def exam_inference(placement_policy, model_name: str):
+@parameterize('model_init_func', [single_chunk_init, multi_chunk_init])
+def exam_inference(placement_policy: str, model_name: str, model_init_func: Callable):
     set_seed(19360226)
     get_components_func = non_distributed_component_funcs.get_callable(model_name)
     model_builder, train_dataloader, test_dataloader, optimizer_class, criterion = get_components_func()
@@ -56,18 +83,7 @@ def exam_inference(placement_policy, model_name: str):
     for torch_p, p in zip(torch_model.parameters(), model.parameters()):
         p.data.copy_(torch_p.data)
 
-    world_size = torch.distributed.get_world_size()
-    config_dict, *_ = search_chunk_configuration(model, search_range_mb=1, search_interval_byte=100)
-    config_dict[world_size]['chunk_size'] = 5000
-    config_dict[world_size]['keep_gathered'] = False
-    if placement_policy != 'cuda':
-        init_device = torch.device('cpu')
-    else:
-        init_device = None
-    chunk_manager = ChunkManager(config_dict, init_device=init_device)
-    gemini_manager = GeminiManager(placement_policy, chunk_manager)
-    model = ZeroDDP(model, gemini_manager, pin_memory=True)
-
+    model = model_init_func(model, placement_policy)
     optimizer = HybridAdam(model.parameters(), lr=1e-3)
     zero_optim = ZeroOptimizer(optimizer, model, initial_scale=128)