[checkpoint] support huggingface style sharded checkpoint (#3461)

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

---------

Co-authored-by: luchen <luchen@luchendeMBP.lan>

colossalai/checkpoint_io/general_checkpoint_io.py

@@ -2,37 +2,35 @@ from pathlib import Path
 
 import torch.nn as nn
 from torch.optim import Optimizer
+import logging
+import os
+import json
+import gc
 
 from .checkpoint_io_base import CheckpointIO
 from .index_file import CheckpointIndexFile
-from .utils import has_index_file, load_state_dict, save_state_dict
+from .utils import (
+    has_index_file, 
+    load_state_dict, 
+    save_state_dict, 
+    is_safetensors_available,
+    shard_checkpoint,
+    load_shard_state_dict,
+    load_state_dict_into_model
+    )
+from .utils import SAFE_WEIGHTS_NAME, WEIGHTS_NAME, SAFE_WEIGHTS_INDEX_NAME, WEIGHTS_INDEX_NAME
 
 __all__ = ['GeneralCheckpointIO']
 
 
 class GeneralCheckpointIO(CheckpointIO):
-
+    """
-    def load_sharded_model(self, model: nn.Module, index_file_path: Path, strict: bool):
+    Checkpoint IO
-        # load the index file
+    """
-        index_file = CheckpointIndexFile.from_file(index_file_path)
-
-        # iterate over the shard checkpoint files
-        # and load each
-        index_file.assert_no_dtensor_checkpoint()
-        checkpoint_file_list, _ = index_file.get_checkpoint_fileanames()
-        for shard_file in checkpoint_file_list:
-            shard_checkpoint = load_state_dict(shard_file)
-            model.load_state_dict(shard_checkpoint, strict=strict)
-
     def load_unsharded_model(self, model: nn.Module, checkpoint: str, strict: bool):
         checkpoint = load_state_dict(checkpoint)
         model.load_state_dict(checkpoint, strict=strict)
 
-    def save_sharded_model(self, model: nn.Module, checkpoint: Path, gather_dtensor: bool, prefix: str,
-                           size_per_shard: int, use_safetensors: bool):
-        # TODO(FrankLeeeee): implement this method as it can be supported by Huggingface model
-        raise NotImplementedError("Sharded model checkpoint is not supported yet.")
-
     def save_unsharded_model(self, model: nn.Module, checkpoint: str, gather_dtensor: bool, use_safetensors: bool):
         state_dict = model.state_dict()
 
@@ -68,3 +66,68 @@ class GeneralCheckpointIO(CheckpointIO):
     ):
         # TODO(FrankLeeeee): handle distributed tensors
         save_state_dict(optimizer.state_dict(), checkpoint, use_safetensors=False)
+
+
+    def save_sharded_model(self, model: nn.Module, checkpoint_path: str, gather_dtensor:bool = False, 
+                           prefix: str = "", max_shard_size: int = 1024, use_safetensors: bool = False):
+        """ 
+        implement this method as it can be supported by Huggingface model,
+        save shard model, save model to multiple files
+        """
+        if os.path.isfile(checkpoint_path):
+            logging.error(f"Provided path ({checkpoint_path}) should be a directory, not a file")
+            return
+        
+        Path(checkpoint_path).mkdir(parents=True, exist_ok=True)
+        
+        # shard checkpoint
+        state_dict = model.state_dict()
+        weights_name = SAFE_WEIGHTS_NAME if use_safetensors else WEIGHTS_NAME
+        shards, index = shard_checkpoint(state_dict, max_shard_size=max_shard_size, weights_name=weights_name)
+
+        # Save the model
+        for shard_file, shard in shards.items():
+            checkpoint_file_path = os.path.join(checkpoint_path, shard_file)
+            save_state_dict(shard, checkpoint_file_path, use_safetensors)
+
+        # save index file
+        save_index_file = SAFE_WEIGHTS_INDEX_NAME if use_safetensors else WEIGHTS_INDEX_NAME
+        save_index_file = os.path.join(checkpoint_path, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+        logging.info(
+            f"The model is bigger than the maximum size per checkpoint ({max_shard_size}) and is going to be "
+            f"split in {len(shards)} checkpoint shards. You can find where each parameters has been saved in the "
+            f"index located at {save_index_file}."
+        )
+
+
+    def load_sharded_model(self, model: nn.Module, checkpoint_index_file: Path, strict: bool = False, use_safetensors: bool = False):
+        """
+        load shard model, load model from multiple files
+        """
+        use_safetensors = False
+        if "safetensors" in checkpoint_index_file.name:
+            use_safetensors = True
+
+        if use_safetensors and not is_safetensors_available():
+            raise ImportError("`safe_serialization` requires the `safetensors` library: `pip install safetensors`.")
+        
+        # read checkpoint index file
+        ckpt_index_file = CheckpointIndexFile.from_file(checkpoint_index_file)
+        checkpoint_files, _ = ckpt_index_file.get_checkpoint_fileanames()
+        missing_keys = ckpt_index_file.get_all_param_names()
+
+        for shard_file in checkpoint_files:
+            state_dict = load_shard_state_dict(Path(shard_file), use_safetensors)
+            load_state_dict_into_model(model, state_dict, missing_keys, strict)
+            del state_dict
+            gc.collect()
+
+        if strict and len(missing_keys) > 0:
+            error_msgs = 'Missing key(s) in state_dict: {}. '.format(
+                        ', '.join('"{}"'.format(k) for k in missing_keys))
+            raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
+                               self.__class__.__name__, "\n\t".join(error_msgs)))
+

colossalai/checkpoint_io/index_file.py

@@ -148,3 +148,9 @@ class CheckpointIndexFile:
         """
         ckpt_path = self.weight_map[param_name]
         return ckpt_path
+    
+    def get_all_param_names(self):
+        """
+        Get all the weight keys.
+        """
+        return list(self.weight_map.keys())

colossalai/checkpoint_io/utils.py

@@ -1,13 +1,19 @@
+# coding=utf-8
 from pathlib import Path
-from typing import List, Optional, Tuple
-
 import torch
+import torch.nn as nn
+from typing import List, Dict, Mapping, OrderedDict, Optional, Tuple
+from colossalai.tensor.d_tensor.d_tensor import DTensor
+
+SAFE_WEIGHTS_NAME = "model.safetensors"
+WEIGHTS_NAME = "model.bin"
+SAFE_WEIGHTS_INDEX_NAME = "model.safetensors.index.json"
+WEIGHTS_INDEX_NAME = "model.bin.index.json"
 
 # ======================================
 # General helper functions
 # ======================================
 
-
 def calculate_tensor_size(tensor: torch.Tensor) -> float:
     """
     Calculate the size of a parameter in MB. Used to compute whether a group of params exceed the shard size.
@@ -68,6 +74,130 @@ def is_safetensor_checkpoint(checkpoint_file_path: str) -> bool:
         return False
 
 
+# ======================================
+# Helper functions for saving shard file
+# ======================================
+def shard_checkpoint(state_dict: torch.Tensor, max_shard_size: int = 1024, weights_name: str = WEIGHTS_NAME):
+ 
+    """
+    Splits a model state dictionary in sub-checkpoints so that the final size of each sub-checkpoint does not exceed a
+    given size.
+    """
+    sharded_state_dicts = []
+    current_block = {}
+    current_block_size = 0
+    total_size = 0
+
+    for key, weight in state_dict.items():
+        if type(weight) != DTensor:
+            weight_size = calculate_tensor_size(weight)
+
+            # If this weight is going to tip up over the maximal size, we split.
+            if current_block_size + weight_size > max_shard_size:
+                sharded_state_dicts.append(current_block)
+                current_block = {}
+                current_block_size = 0
+
+            current_block[key] = weight
+            current_block_size += weight_size
+            total_size += weight_size
+
+    # Add the last block
+    sharded_state_dicts.append(current_block)
+
+    # If we only have one shard, we return it
+    if len(sharded_state_dicts) == 1:
+        return {weights_name: sharded_state_dicts[0]}, None
+    
+    # Otherwise, let's build the index
+    weight_map = {}
+    shards = {}
+
+    for idx, shard in enumerate(sharded_state_dicts):
+        shard_file = weights_name.replace(".bin", f"-{idx+1:05d}-of-{len(sharded_state_dicts):05d}.bin")
+        shard_file = shard_file.replace(
+            ".safetensors", f"-{idx + 1:05d}-of-{len(sharded_state_dicts):05d}.safetensors"
+        )
+        shards[shard_file] = shard
+        for key in shard.keys():
+            weight_map[key] = shard_file
+
+    # Add the metadata
+    metadata = {"total_size": total_size}
+    index = {"metadata": metadata, "weight_map": weight_map}
+    return shards, index
+
+def load_shard_state_dict(checkpoint_file: Path, use_safetensors: bool =False):
+    """
+    load shard state dict into model
+    """
+    if use_safetensors and not checkpoint_file.suffix == ".safetensors":
+        raise Exception("load the model using `safetensors`, but no file endwith .safetensors")
+    if use_safetensors:
+        from safetensors.torch import safe_open
+        from safetensors.torch import load_file as safe_load_file
+        with safe_open(checkpoint_file, framework="pt") as f:
+            metadata = f.metadata()
+        if metadata["format"] != "pt":
+            raise NotImplementedError(
+                f"Conversion from a {metadata['format']} safetensors archive to PyTorch is not implemented yet."
+            )
+        return safe_load_file(checkpoint_file)
+    else:
+        return torch.load(checkpoint_file)
+    
+def load_state_dict_into_model(model: nn.Module, state_dict: torch.Tensor, missing_keys: List, strict: bool = False):
+    r"""Copies parameters and buffers from :attr:`state_dict` into
+    this module and its descendants. 
+
+    Args:
+        state_dict (dict): a dict containing parameters and
+            persistent buffers.
+    """
+    if not isinstance(state_dict, Mapping):
+        raise TypeError("Expected state_dict to be dict-like, got {}.".format(type(state_dict)))
+
+    unexpected_keys: List[str] = []
+    sub_missing_keys: List[str] = []
+    error_msgs: List[str] = []
+
+    # copy state_dict so _load_from_state_dict can modify it
+    metadata = getattr(state_dict, '_metadata', None)
+    state_dict = OrderedDict(state_dict)
+    if metadata is not None:
+        state_dict._metadata = metadata
+
+    def load(module: nn.Module, state_dict, prefix=""):
+        local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
+        args = (state_dict, prefix, local_metadata, True, [], [], error_msgs)
+        # Parameters of module and children will start with prefix. We can exit early if there are none in this
+        # state_dict
+        if len([key for key in state_dict if key.startswith(prefix)]) > 0:
+            module._load_from_state_dict(*args)
+
+        for name, child in module._modules.items():
+            if child is not None:
+                load(child, state_dict, prefix + name + ".")
+
+    load(model, state_dict, "")
+    del load
+
+    # deal with missing key
+    if len(missing_keys) > 0:
+        deleted_keys = []
+        for key in missing_keys:
+            if key not in sub_missing_keys:
+                deleted_keys.append(key)
+        for key in deleted_keys:
+            missing_keys.remove(key)
+
+    if strict:
+        if len(unexpected_keys) > 0:
+            error_msgs = 'Unexpected key(s) in state_dict: {}. '.format(
+                        ', '.join('"{}"'.format(k) for k in unexpected_keys))
+            raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
+                               model.__class__.__name__, "\n\t".join(error_msgs)))
+        
 # ======================================
 # Helper functions for saving state dict
 # ======================================
@@ -86,8 +216,8 @@ def save_state_dict(state_dict: dict, checkpoint_file_path: str, use_safetensors
         assert is_safetensors_available(), "safetensors is not available."
         assert checkpoint_file_path.endswith('.safetensors'), \
             "safetensors only supports .safetensors suffix for checkpoint file."
-        from safetensors.torch import save_file
+        from safetensors.torch import save_file as safe_save_file
-        save_file(state_dict, checkpoint_file_path)
+        safe_save_file(state_dict, checkpoint_file_path, metadata={"format": "pt"})
     else:
         torch.save(state_dict, checkpoint_file_path)
 

tests/test_checkpoint_io/test_general_checkpoint_io.py

@@ -1,9 +1,12 @@
 import tempfile
-
 import pytest
 import torch
+import logging
 from torch.optim import Adam
 from torchvision.models import resnet18
+from pathlib import Path
+import os
+import subprocess
 
 from colossalai.checkpoint_io import GeneralCheckpointIO
 from colossalai.testing import clear_cache_before_run, parameterize
@@ -12,7 +15,7 @@ from colossalai.testing import clear_cache_before_run, parameterize
 # Note:
 # 1. due to checkpoint IO can be quite slow if tested with all models, we will only test on resnet for now
 # 2. we will test on both sharded and unsharded checkpoints
-# 3. TODO(FrankLeeeee): implement sharded checkpoint and test it
+# 3. implement sharded checkpoint and test it
 # ========
 
 
@@ -53,12 +56,61 @@ def test_unsharded_checkpoint(use_safetensors: bool):
     ckpt_io.load_model(new_model, model_ckpt_tempfile.name)
     ckpt_io.load_optimizer(new_optimizer, optimizer_ckpt_tempfile.name)
 
-    # do recursive check for the optimizer state dict
+
-    # if the value is a dict, compare its values
+    # check for model and optimizer state dict recursively
-    # if the value is a list, comapre all elements one-by-one
+    recursive_check(model.state_dict(), new_model.state_dict())
-    # if the value is a torch.Tensor, use torch.equal
+    recursive_check(optimizer.state_dict(), new_optimizer.state_dict())
-    # otherwise use assertEqual
+
-    def recursive_check(d1, d2):
+@pytest.mark.parametrize('use_safetensors', [True, False])
+def test_sharded_checkpoint(use_safetensors: bool):
+    # create a model and optimizer
+    model = resnet18()
+    optimizer = Adam(model.parameters(), lr=0.001)
+    # create test data sample
+    x = torch.randn(1, 3, 224, 224)
+
+    # run fwd and bwd
+    y = model(x)
+    loss = y.sum()
+    loss.backward()
+    optimizer.step()
+
+    # create a temp file for checkpoint
+    if use_safetensors:
+        suffix = ".safetensors"
+        SAFE_WEIGHTS_INDEX_NAME = "model.safetensors.index.json"
+    else:
+        suffix = ".bin"
+        WEIGHTS_INDEX_NAME = "model.bin.index.json"
+    
+    # model_ckpt_dir = tempfile.TemporaryDirectory(suffix=suffix)
+    model_ckpt_dir = tempfile.TemporaryDirectory()
+    optimizer_ckpt_tempfile = tempfile.NamedTemporaryFile()
+
+    # save the model and optimizer
+    ckpt_io = GeneralCheckpointIO()
+
+    ckpt_io.save_model(model, model_ckpt_dir.name, True, True, "", 10, use_safetensors=use_safetensors)
+    ckpt_io.save_optimizer(optimizer, optimizer_ckpt_tempfile.name, shard=False)
+    
+    # create new model
+    new_model = resnet18()
+    new_optimizer = Adam(new_model.parameters(), lr=0.001)
+
+    ckpt_io.load_model(new_model, str(model_ckpt_dir.name), strict=True)
+    ckpt_io.load_optimizer(new_optimizer, optimizer_ckpt_tempfile.name)
+
+    # check for model and optimizer state dict recursively
+    recursive_check(model.state_dict(), new_model.state_dict())
+    recursive_check(optimizer.state_dict(), new_optimizer.state_dict())
+
+
+# do recursive check for the optimizer state dict
+# if the value is a dict, compare its values
+# if the value is a list, comapre all elements one-by-one
+# if the value is a torch.Tensor, use torch.equal
+# otherwise use assertEqual
+def recursive_check(d1, d2):
     for k, v in d1.items():
         if isinstance(v, dict):
             recursive_check(v, d2[k])
@@ -72,8 +124,3 @@ def test_unsharded_checkpoint(use_safetensors: bool):
             assert torch.equal(v, d2[k])
         else:
             assert v == d2[k]
-
-    # check for model and optimizer state dict recursively
-    recursive_check(model.state_dict(), new_model.state_dict())
-    recursive_check(optimizer.state_dict(), new_optimizer.state_dict())
-    recursive_check(optimizer.state_dict(), new_optimizer.state_dict())