[ckpt] Add async ckpt api for other plugin

colossalai/booster/plugin/gemini_plugin.py

@@ -82,7 +82,7 @@ def save_unsharded_model(
         state_dict = model.state_dict(only_rank_0=True)
         if self.coordinator.is_master():
             if use_async:
-                super().save_unsharded_model(model, checkpoint, gather_dtensor, use_safetensors, use_async)
+                super().save_unsharded_model(model, checkpoint, gather_dtensor, use_safetensors, use_async, state_dict)
             else:
                 save_state_dict(state_dict, checkpoint, use_safetensors)
 

colossalai/checkpoint_io/general_checkpoint_io.py

@@ -45,11 +45,16 @@ def load_unsharded_model(self, model: nn.Module, checkpoint: str, strict: bool):
         model.load_state_dict(checkpoint, strict=strict)
 
     def save_unsharded_model(
-        self, model: nn.Module, checkpoint: str, gather_dtensor: bool, use_safetensors: bool, use_async: bool = False
+        self,
+        model: nn.Module,
+        checkpoint: str,
+        gather_dtensor: bool,
+        use_safetensors: bool,
+        use_async: bool = False,
+        state_dict: dict = None,
     ):
-        state_dict = model.state_dict()
-
-        # TODO(FrankLeeeee): add support for gather_dtensor
+        if state_dict is None:
+            state_dict = model.state_dict()
         if gather_dtensor:
             pass
 
@@ -60,11 +65,11 @@ def save_unsharded_model(
             if id(model) not in self.pinned_state_dicts:
                 self.pinned_state_dicts[id(model)] = create_pinned_state_dict(state_dict)
             self.async_writers.append(writer)
-            move_and_save(writer, state_dict, self.pinned_state_dicts[id(model)])
+            move_and_save(writer, state_dict, state_dict_pinned=self.pinned_state_dicts[id(model)])
 
         else:
             # save the checkpoint
-            save_state_dict(state_dict, checkpoint, use_safetensors)
+            save_state_dict(model.state_dict(), checkpoint, use_safetensors)
 
     def load_sharded_optimizer(self, optimizer: Optimizer, index_file_path: str, prefix: str):
         """

colossalai/checkpoint_io/hybrid_parallel_checkpoint_io.py

@@ -9,6 +9,7 @@
 import torch
 import torch.distributed as dist
 import torch.nn as nn
+from tensornvme.async_file_io import AsyncFileWriter
 from torch.distributed import ProcessGroup
 from torch.optim.lr_scheduler import _LRScheduler as LRScheduler
 from torch.utils._pytree import tree_map
@@ -22,6 +23,7 @@
     to_unpadded_tensor,
 )
 from colossalai.utils import get_current_device, get_non_persistent_buffers_set
+from colossalai.utils.safetensors import save
 
 from .general_checkpoint_io import GeneralCheckpointIO
 from .index_file import CheckpointIndexFile
@@ -199,7 +201,6 @@ def save_sharded_model(
             use_safetensors (bool, optional): Whether to use safe tensors. Defaults to False.
             use_async (bool, optional): Whether to save the state_dicts of model asynchronously. Defaults to False.
         """
-
         assert isinstance(model, ModelWrapper), "Please boost the model before saving!"
         model._force_wait_all_gather()
         model = model.unwrap()
@@ -224,7 +225,18 @@ def save_sharded_model(
         if self.pp_size == 1:
             # When pipeline is not used, save the model shards as in general checkpointIO
             if use_async:
-                super().save_unsharded_model(model, checkpoint, gather_dtensor, use_safetensors, use_async=use_async)
+                pinned_state_dict = self.pinned_state_dicts.get(id(model), None)
+                total_size, new_pinned_state_dict, writers = async_save_state_dict_shards(
+                    sharded_state_dict=state_dict_shard,
+                    checkpoint=checkpoint,
+                    index_file=index_file,
+                    base_filename=weights_name,
+                    is_master=control_saving,
+                    pinned_state_dict=pinned_state_dict,
+                    n_write_entries=self.N_WRITE_ENTRIES,
+                )
+                self.pinned_state_dicts[id(model)] = new_pinned_state_dict
+                self.async_writers.extend(writers)
             else:
                 total_size = save_state_dict_shards(
                     sharded_state_dict=state_dict_shard,
@@ -234,16 +246,16 @@ def save_sharded_model(
                     is_master=control_saving,
                     use_safetensors=use_safetensors,
                 )
-                if control_saving:
-                    index_file.append_meta_data("total_size", total_size)
-                    index_file.write_index_file(save_index_file)
-                    save_config_file(model, checkpoint)
-                    if self.verbose and self.coordinator.is_master():
-                        logging.info(
-                            f"The model is split into checkpoint shards. "
-                            f"You can find where each parameters has been saved in the "
-                            f"index located at {save_index_file}."
-                        )
+            if control_saving:
+                index_file.append_meta_data("total_size", total_size)
+                index_file.write_index_file(save_index_file)
+                save_config_file(model, checkpoint)
+                if self.verbose and self.coordinator.is_master():
+                    logging.info(
+                        f"The model is split into checkpoint shards. "
+                        f"You can find where each parameters has been saved in the "
+                        f"index located at {save_index_file}."
+                    )
 
         else:
             # When pipeline is used, each stage produces its own shard files and index files.
@@ -259,24 +271,30 @@ def save_sharded_model(
             save_index_file = save_index_file.replace(".json", f"-stage-{self.pp_rank+1:05d}.json")
             save_index_file = os.path.join("tmp_index_files", save_index_file)
             if use_async:
-                total_size, returned_state_dict, writers = async_save_state_dict_shards(
+                pinned_state_dict = self.pinned_state_dicts.get(id(model), None)
+                total_size, new_pinned_state_dict, writers = async_save_state_dict_shards(
                     sharded_state_dict=state_dict_shard,
                     checkpoint=checkpoint,
                     index_file=index_file,
                     base_filename=weights_name,
                     is_master=control_saving,
                     use_pp_format=True,
-                    n_write_entries=191,
+                    pinned_state_dict=pinned_state_dict,
+                    n_write_entries=self.N_WRITE_ENTRIES,
+                    move=False,
+                )
+                self.pinned_state_dicts[id(model)] = new_pinned_state_dict
+                self.async_writers.extend(writers)
+            else:
+                total_size = save_state_dict_shards(
+                    sharded_state_dict=state_dict_shard,
+                    checkpoint=checkpoint,
+                    index_file=index_file,
+                    base_filename=weights_name,
+                    is_master=control_saving,
+                    use_safetensors=use_safetensors,
+                    use_pp_format=True,
                 )
-            total_size = save_state_dict_shards(
-                sharded_state_dict=state_dict_shard,
-                checkpoint=checkpoint,
-                index_file=index_file,
-                base_filename=weights_name,
-                is_master=control_saving,
-                use_safetensors=use_safetensors,
-                use_pp_format=True,
-            )
 
             if control_saving:
                 assert (
@@ -664,7 +682,6 @@ def save_unsharded_model(
         model = model.unwrap()
         if self.dp_rank != 0:
             return
-
         # The logic of collecting parameter shards along tp degree
         # has been implemented by _save_to_state_dict method of ParallelModule in Shardformer.
         state_dict = model.state_dict()
@@ -686,15 +703,12 @@ def save_unsharded_model(
                 for _state_dict in state_dict_list:
                     complete_state_dict.update(_state_dict)
                 if use_async:
-                    from tensornvme.async_file_io import AsyncFileWriter
-
-                    from colossalai.utils.safetensors import move_and_save
 
                     writer = AsyncFileWriter(checkpoint, self.N_WRITE_ENTRIES, backend="pthread")
                     if id(model) not in self.pinned_state_dicts:
-                        self.pinned_state_dicts[id(model)] = create_pinned_state_dict(state_dict)
+                        self.pinned_state_dicts[id(model)] = create_pinned_state_dict(complete_state_dict)
                     self.async_writers.append(writer)
-                    move_and_save(writer, state_dict, self.pinned_state_dicts[id(model)])
+                    save(writer, complete_state_dict, self.pinned_state_dicts[id(model)])
                 else:
                     save_state_dict(complete_state_dict, checkpoint, use_safetensors)
 

colossalai/checkpoint_io/utils.py

@@ -321,7 +321,7 @@ def async_save_state_dict_shards(
             returned_state_dict.update(sub_pinned_state_dict)
 
         # Only save on master rank.
-        move_and_save(writer, shard, sub_pinned_state_dict)
+        move_and_save(writer, shard, state_dict_pinned=sub_pinned_state_dict)
         shard_filenames.append(shard_file)
         del shard
 

colossalai/utils/safetensors.py

@@ -170,9 +170,10 @@ def save_nested(f_writer: AsyncFileWriter, state_dict: Dict[str, torch.Tensor])
 def move_and_save(
     f_writer: AsyncFileWriter,
     state_dict: Dict[str, torch.Tensor],
+    metadata: Optional[Dict[str, str]] = None,
     state_dict_pinned: Optional[Dict[str, torch.Tensor]] = None,
 ) -> None:
-    prepared_data, _, tensor_keys = prepare(state_dict)
+    prepared_data, _, tensor_keys = prepare(state_dict, metadata)
     n, header_bytes, _ = prepared_data.n, prepared_data.header_bytes, prepared_data.offset
 
     f_writer.write(n.to_bytes(8, byteorder="little"))

tests/test_checkpoint_io/test_gemini_checkpoint_io.py

@@ -35,7 +35,13 @@
 @parameterize("use_safetensors", [False, True])
 @parameterize("tp_size", [1, 2])
 @parameterize("zero_size", [2])
-def exam_state_dict_with_origin(placement_config, model_name, use_safetensors: bool, tp_size: int, zero_size: int):
+def exam_state_dict_with_origin(
+    placement_config,
+    model_name,
+    use_safetensors: bool,
+    tp_size: int,
+    zero_size: int,
+):
     from transformers import BertForSequenceClassification
 
     (model_fn, data_gen_fn, output_transform_fn, _, _) = next(iter(model_zoo.get_sub_registry(model_name).values()))
@@ -71,19 +77,24 @@ def exam_state_dict_with_origin(placement_config, model_name, use_safetensors: b
             (model_size / 3),
             use_safetensors=use_safetensors,
         )
+        booster.checkpoint_io._sync_d2h()
+        booster.checkpoint_io._sync_io()
         dist.barrier()
         new_bert_model = BertForSequenceClassification.from_pretrained(pretrained_path)
         check_state_dict_equal(bert_model.state_dict(only_rank_0=False), new_bert_model.state_dict())
 
 
 @clear_cache_before_run()
 @parameterize("placement_config", OPTIM_PLACEMENT_CONFIGS)
-@parameterize("shard", [True, False])
+@parameterize("shard", [False, True])
 @parameterize("model_name", ["transformers_llama_for_causal_lm"])
 @parameterize("size_per_shard", [32])
 @parameterize("tp_size", [1, 2])
 @parameterize("zero_size", [2])
-def exam_state_dict(placement_config, shard: bool, model_name: str, size_per_shard: int, tp_size: int, zero_size: int):
+@parameterize("use_async", [False, True])
+def exam_state_dict(
+    placement_config, shard: bool, model_name: str, size_per_shard: int, tp_size: int, zero_size: int, use_async: bool
+):
     (model_fn, data_gen_fn, output_transform_fn, _, _) = next(iter(model_zoo.get_sub_registry(model_name).values()))
     criterion = lambda x: x.mean()
     enable_flash_attention = True if tp_size > 1 else False
@@ -124,14 +135,13 @@ def exam_state_dict(placement_config, shard: bool, model_name: str, size_per_sha
     with shared_tempdir() as tempdir:
         model_ckpt_path = f"{tempdir}/model"
         optimizer_ckpt_path = f"{tempdir}/optimizer"
-        booster.save_model(
-            model,
-            model_ckpt_path,
-            shard=shard,
-            size_per_shard=size_per_shard,
-        )
+        if not shard and use_async:
+            model_ckpt_path = f"{model_ckpt_path}.safetensors"
+        booster.save_model(model, model_ckpt_path, shard=shard, size_per_shard=size_per_shard, use_async=use_async)
 
         booster.save_optimizer(optimizer, optimizer_ckpt_path, shard=shard, size_per_shard=size_per_shard)
+        booster.checkpoint_io._sync_d2h()
+        booster.checkpoint_io._sync_io()
         dist.barrier()
 
         booster.load_model(new_model, model_ckpt_path)

tests/test_checkpoint_io/test_hybrid_parallel_plugin_checkpoint_io.py

@@ -34,6 +34,7 @@
 else:
     TEST_CONFIGS = [
         # TODO(ver217): other configs lead to hang
+        {"tp_size": 1, "pp_size": 1, "num_microbatches": 4, "zero_stage": 1, "precision": "fp16", "initial_scale": 1},
         {"tp_size": 1, "pp_size": 2, "num_microbatches": 4, "zero_stage": 1, "precision": "fp16", "initial_scale": 1},
     ]
 
@@ -42,8 +43,9 @@
 @parameterize("model_name", ["transformers_llama_for_causal_lm"])
 @parameterize("size_per_shard", [32])
 @parameterize("test_config", TEST_CONFIGS)
+@parameterize("use_async", [True, False])
 @clear_cache_before_run()
-def exam_state_dict(shard: bool, model_name: str, size_per_shard: int, test_config: dict):
+def exam_state_dict(shard: bool, model_name: str, size_per_shard: int, test_config: dict, use_async: bool):
     (model_fn, data_gen_fn, output_transform_fn, loss_fn, _) = next(
         iter(model_zoo.get_sub_registry(model_name).values())
     )
@@ -85,8 +87,12 @@ def _preprocess_data(data):
     with shared_tempdir() as tempdir:
         model_ckpt_path = f"{tempdir}/model"
         optimizer_ckpt_path = f"{tempdir}/optimizer"
-        booster.save_model(model, model_ckpt_path, shard=shard, size_per_shard=size_per_shard)
+        if not shard and use_async:
+            model_ckpt_path = f"{model_ckpt_path}.safetensors"
+        booster.save_model(model, model_ckpt_path, shard=shard, size_per_shard=size_per_shard, use_async=use_async)
         booster.save_optimizer(optimizer, optimizer_ckpt_path, shard=shard, size_per_shard=size_per_shard)
+        booster.checkpoint_io._sync_d2h()
+        booster.checkpoint_io._sync_io()
         dist.barrier()
 
         new_model = model_fn().cuda()

tests/test_checkpoint_io/test_low_level_zero_checkpoint_io.py

@@ -47,8 +47,6 @@ def check_low_level_zero_checkpointIO(stage: int, shard: bool, offload: bool, us
 
         model_ckpt_path = f"{tempdir}/model"
         optimizer_ckpt_path = f"{tempdir}/optimizer"
-        if not shard and not use_async:
-            model_ckpt_path = f"{model_ckpt_path}.pt"
         if not shard and use_async:
             model_ckpt_path = f"{model_ckpt_path}.safetensors"
         if not shard and use_async:

tests/test_checkpoint_io/test_safetensors_async_io.py

@@ -146,7 +146,11 @@ def test_save_load():
         model_state_pinned = {k: v.pin_memory() for k, v in model_state_dict.items()}
         model_saved_path = f"{tempdir}/save_model_cuda.safetensors"
         f_writer = AsyncFileWriter(model_saved_path, n_entries=191, backend="pthread")
-        move_and_save(f_writer, model_state_dict_cuda, model_state_pinned)
+        move_and_save(
+            f_writer,
+            state_dict=model_state_dict_cuda,
+            state_dict_pinned=model_state_pinned,
+        )
         f_writer.sync_before_step()
         f_writer.synchronize()
         del f_writer

tests/test_checkpoint_io/test_torch_ddp_checkpoint_io.py

@@ -13,8 +13,9 @@
 
 
 @parameterize("shard", [True, False])
+@parameterize("use_async", [True, False])
 @parameterize("size_per_shard", [16, 128])
-def check_torch_ddp_checkpointIO(shard: bool, size_per_shard: int):
+def check_torch_ddp_checkpointIO(shard: bool, use_async: bool, size_per_shard: int):
     plugin = TorchDDPPlugin()
     booster = Booster(plugin=plugin)
     model = resnet18()
@@ -39,7 +40,11 @@ def check_torch_ddp_checkpointIO(shard: bool, size_per_shard: int):
         model_ckpt_path = f"{tempdir}/model"
         optimizer_ckpt_path = f"{tempdir}/optimizer"
         lr_scheduler_ckpt_path = f"{tempdir}/lr_scheduler"
-        booster.save_model(model, model_ckpt_path, shard=shard, size_per_shard=size_per_shard)
+        if not use_async:
+            model_ckpt_path = f"{model_ckpt_path}.pt"
+        if use_async:
+            model_ckpt_path = f"{model_ckpt_path}.safetensors"
+        booster.save_model(model, model_ckpt_path, shard=shard, size_per_shard=size_per_shard, use_async=use_async)
         booster.save_optimizer(optimizer, optimizer_ckpt_path, shard=shard, size_per_shard=size_per_shard)
         booster.save_lr_scheduler(scheduler, lr_scheduler_ckpt_path)
         dist.barrier()