预训练全参数训练过程中checkpoint参数保存为空

[Double-bear]

提交前必须检查以下项目

• [X] 请确保使用的是仓库最新代码（git pull），一些问题已被解决和修复。
• [X] 我已阅读项目文档和FAQ章节并且已在Issue中对问题进行了搜索，没有找到相似问题和解决方案。
• [X] 第三方插件问题：例如llama.cpp、LangChain、text-generation-webui等，同时建议到对应的项目中查找解决方案。

问题类型

模型训练与精调

基础模型

Chinese-LLaMA-2 (7B/13B)

操作系统

Linux

详细描述问题

我将run_clm_pt_with_peft.py中关于lora的部分删除了，想要进行全参数的训练，发现通过callback函数保存的checkpoint的bin文件非常小，请问这个有什么解决方案吗？ 预训练代码：

#!/usr/bin/env python
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. 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.
"""
Fine-tuning the library models for causal language modeling (GPT, GPT-2, CTRL, ...) on a text file or a dataset.

Here is the full list of checkpoints on the hub that can be fine-tuned by this script:
https://huggingface.co/models?filter=text-generation
"""
# You can also adapt this script on your own causal language modeling task. Pointers for this are left as comments.

import logging
import numpy as np
import math
import os
import torch.distributed as dist
import sys
from dataclasses import dataclass, field
from itertools import chain
from typing import Optional, List, Dict, Any, Mapping
from pathlib import Path
import datasets
import torch
from datasets import load_dataset, concatenate_datasets
import wandb
wandb.login(key="0b8601f0acadfd20eea3bcee89e004fe2d0cf9e7")

import transformers
from transformers import (
    CONFIG_MAPPING,
    MODEL_FOR_CAUSAL_LM_MAPPING,
    AutoConfig,
    AutoModelForCausalLM,
    LlamaForCausalLM,
    LlamaTokenizer,
    AutoTokenizer,
    HfArgumentParser,
    Trainer,
    TrainingArguments,
    is_torch_tpu_available,
    set_seed,
)
from transformers.testing_utils import CaptureLogger
from transformers.trainer_utils import get_last_checkpoint
from transformers.utils import send_example_telemetry
from transformers.utils.versions import require_version

from sklearn.metrics import accuracy_score
from peft import LoraConfig, TaskType, get_peft_model, PeftModel, get_peft_model_state_dict
from transformers.trainer_utils import PREFIX_CHECKPOINT_DIR

class SaveModelCallback(transformers.TrainerCallback):
    def save_model(self, args, state, model, tokenizer, kwargs):
        if state.best_model_checkpoint is not None:
            checkpoint_folder = os.path.join(state.best_model_checkpoint, "pt_model")
        else:
            checkpoint_folder = os.path.join(args.output_dir, f"{PREFIX_CHECKPOINT_DIR}-{state.global_step}")

        pt_model_path = os.path.join(checkpoint_folder, "pt_lora_model")
        model.save_pretrained(pt_model_path)
        tokenizer.save_pretrained(pt_model_path)

    def on_save(self, args, state, control, model, tokenizer, **kwargs):
        self.save_model(args, state, model, tokenizer, kwargs)
        return control

    def on_train_end(self, args, state, control, model, tokenizer, **kwargs):
        pt_model_path = os.path.join(args.output_dir, "pt_model")
        model.save_pretrained(pt_model_path)
        tokenizer.save_pretrained(pt_model_path)

# class CustomTrainer(Trainer):
#     def training_step(self, model, inputs):
#         # 调用原始的training_step
#         loss = super().training_step(model, inputs)

#         # 每N个步骤保存模型
#         if self.state.global_step % 20 == 0:
#             if self.state.best_model_checkpoint is not None:
#                 checkpoint_folder = os.path.join(self.state.best_model_checkpoint, "pt_model")
#             else:
#                 checkpoint_folder = os.path.join(self.args.output_dir, f"{PREFIX_CHECKPOINT_DIR}-{self.state.global_step}")
#             pt_model_path = os.path.join(checkpoint_folder, "pt_lora_model")
#             model.save_pretrained(pt_model_path)
#             if self.tokenizer is not None:
#                 self.tokenizer.save_pretrained(checkpoint_folder)

#         return loss

def accuracy(predictions, references, normalize=True, sample_weight=None):
        return {
            "accuracy": float(
                accuracy_score(references, predictions, normalize=normalize, sample_weight=sample_weight)
            )
        }

def compute_metrics(eval_preds):
    preds, labels = eval_preds
    # preds have the same shape as the labels, after the argmax(-1) has been calculated
    # by preprocess_logits_for_metrics but we need to shift the labels
    labels = labels[:, 1:].reshape(-1)
    preds = preds[:, :-1].reshape(-1)
    return accuracy(predictions=preds, references=labels)

def preprocess_logits_for_metrics(logits, labels):
    if isinstance(logits, tuple):
        # Depending on the model and config, logits may contain extra tensors,
        # like past_key_values, but logits always come first
        logits = logits[0]
    return logits.argmax(dim=-1)

def fault_tolerance_data_collator(features: List) -> Dict[str, Any]:
    if not isinstance(features[0], Mapping):
        features = [vars(f) for f in features]
    first = features[0]
    batch = {}

    # Special handling for labels.
    # Ensure that tensor is created with the correct type
    # (it should be automatically the case, but let's make sure of it.)
    if "label" in first and first["label"] is not None:
        label = first["label"].item() if isinstance(first["label"], torch.Tensor) else first["label"]
        dtype = torch.long if isinstance(label, int) else torch.float
        batch["labels"] = torch.tensor([f["label"] for f in features], dtype=dtype)
    elif "label_ids" in first and first["label_ids"] is not None:
        if isinstance(first["label_ids"], torch.Tensor):
            batch["labels"] = torch.stack([f["label_ids"] for f in features])
        else:
            dtype = torch.long if isinstance(first["label_ids"][0], int) else torch.float
            batch["labels"] = torch.tensor([f["label_ids"] for f in features], dtype=dtype)

    # Handling of all other possible keys.
    # Again, we will use the first element to figure out which key/values are not None for this model.

    try:
        for k, v in first.items():
            if k not in ("label", "label_ids") and v is not None and not isinstance(v, str):
                if isinstance(v, torch.Tensor):
                    batch[k] = torch.stack([f[k] for f in features])
                elif isinstance(v, np.ndarray):
                    batch[k] = torch.tensor(np.stack([f[k] for f in features]))
                else:
                    batch[k] = torch.tensor([f[k] for f in features])
    except ValueError: # quick fix by simply take the first example
        for k, v in first.items():
            if k not in ("label", "label_ids") and v is not None and not isinstance(v, str):
                if isinstance(v, torch.Tensor):
                    batch[k] = torch.stack([features[0][k]] * len(features))
                elif isinstance(v, np.ndarray):
                    batch[k] = torch.tensor(np.stack([features[0][k]] * len(features)))
                else:
                    batch[k] = torch.tensor([features[0][k]] * len(features))

    return batch

MODEL_CONFIG_CLASSES = list(MODEL_FOR_CAUSAL_LM_MAPPING.keys())
MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)

@dataclass
class ModelArguments:
    """
    Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
    """

    model_name_or_path: Optional[str] = field(
        default=None,
        metadata={
            "help": (
                "The model checkpoint for weights initialization.Don't set if you want to train a model from scratch."
            )
        },
    )
    tokenizer_name_or_path: Optional[str] = field(
        default=None,
        metadata={
            "help": (
                "The tokenizer for weights initialization.Don't set if you want to train a model from scratch."
            )
        },
    )
    model_type: Optional[str] = field(
        default=None,
        metadata={"help": "If training from scratch, pass a model type from the list: " + ", ".join(MODEL_TYPES)},
    )
    config_overrides: Optional[str] = field(
        default=None,
        metadata={
            "help": (
                "Override some existing default config settings when a model is trained from scratch. Example: "
                "n_embd=10,resid_pdrop=0.2,scale_attn_weights=false,summary_type=cls_index"
            )
        },
    )
    config_name: Optional[str] = field(
        default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
    )
    tokenizer_name: Optional[str] = field(
        default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
    )
    cache_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
    )
    use_fast_tokenizer: bool = field(
        default=True,
        metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
    )
    model_revision: str = field(
        default="main",
        metadata={"help": "The specific model version to use (can be a branch name, tag name or commit id)."},
    )
    use_auth_token: bool = field(
        default=False,
        metadata={
            "help": (
                "Will use the token generated when running `huggingface-cli login` (necessary to use this script "
                "with private models)."
            )
        },
    )
    torch_dtype: Optional[str] = field(
        default=None,
        metadata={
            "help": (
                "Override the default `torch.dtype` and load the model under this dtype. If `auto` is passed, the "
                "dtype will be automatically derived from the model's weights."
            ),
            "choices": ["auto", "bfloat16", "float16", "float32"],
        },
    )

    def __post_init__(self):
        if self.config_overrides is not None and (self.config_name is not None or self.model_name_or_path is not None):
            raise ValueError(
                "--config_overrides can't be used in combination with --config_name or --model_name_or_path"
            )

@dataclass
class DataTrainingArguments:
    """
    Arguments pertaining to what data we are going to input our model for training and eval.
    """

    dataset_dir: Optional[str] = field(
        default=None, metadata={"help": "The name of the dataset to use (via the datasets library)."}
    )
    dataset_config_name: Optional[str] = field(
        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
    )
    train_file: Optional[str] = field(default=None, metadata={"help": "The input training data file (a text file)."})
    validation_file: Optional[str] = field(
        default=None,
        metadata={"help": "An optional input evaluation data file to evaluate the perplexity on (a text file)."},
    )
    max_train_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": (
                "For debugging purposes or quicker training, truncate the number of training examples to this "
                "value if set."
            )
        },
    )
    max_eval_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": (
                "For debugging purposes or quicker training, truncate the number of evaluation examples to this "
                "value if set."
            )
        },
    )
    streaming: bool = field(default=False, metadata={"help": "Enable streaming mode"})
    block_size: Optional[int] = field(
        default=None,
        metadata={
            "help": (
                "Optional input sequence length after tokenization. "
                "The training dataset will be truncated in block of this size for training. "
                "Default to the model max input length for single sentence inputs (take into account special tokens)."
            )
        },
    )
    overwrite_cache: bool = field(
        default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
    )
    validation_split_percentage: Optional[float] = field(
        default=0.05,
        metadata={
            "help": "The percentage of the train set used as validation set in case there's no validation split"
        },
    )
    preprocessing_num_workers: Optional[int] = field(
        default=None,
        metadata={"help": "The number of processes to use for the preprocessing."},
    )
    keep_linebreaks: bool = field(
        default=True, metadata={"help": "Whether to keep line breaks when using TXT files or not."}
    )
    data_cache_dir: Optional[str] = field(default="./", metadata={"help": "The datasets processed stored"})

    def __post_init__(self):
        if self.streaming:
            require_version("datasets>=2.0.0", "The streaming feature requires `datasets>=2.0.0`")

@dataclass
class MyTrainingArguments(TrainingArguments):
    trainable : Optional[str] = field(default="q_proj,v_proj")
    lora_rank : Optional[int] = field(default=8)
    lora_dropout : Optional[float] = field(default=0.1)
    lora_alpha : Optional[float] = field(default=32.)
    modules_to_save : Optional[str] = field(default=None)
    debug_mode : Optional[bool] = field(default=False)
    flash_attn : Optional[bool] = field(default=False)

logger = logging.getLogger(__name__)

def main():

    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, MyTrainingArguments))
    if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
        # If we pass only one argument to the script and it's the path to a json file,
        # let's parse it to get our arguments.
        model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
    else:
        model_args, data_args, training_args = parser.parse_args_into_dataclasses()
    if training_args.flash_attn:
        from flash_attn_patch import replace_llama_attn_with_flash_attn
        replace_llama_attn_with_flash_attn()

    # Sending telemetry. Tracking the example usage helps us better allocate resources to maintain them. The
    # information sent is the one passed as arguments along with your Python/PyTorch versions.
    send_example_telemetry("run_clm", model_args, data_args)

    # Setup logging
    logging.basicConfig(format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",datefmt="%m/%d/%Y %H:%M:%S",
        level=logging.INFO,  # if training_args.local_rank in [-1, 0] else logging.WARN,
        handlers=[logging.StreamHandler(sys.stdout)],)

    if training_args.should_log:
        # The default of training_args.log_level is passive, so we set log level at info here to have that default.
        transformers.utils.logging.set_verbosity_info()

    log_level = training_args.get_process_log_level()
    logger.setLevel(log_level)
    datasets.utils.logging.set_verbosity(log_level)
    transformers.utils.logging.set_verbosity(log_level)
    transformers.utils.logging.enable_default_handler()
    transformers.utils.logging.enable_explicit_format()
    # transformers.tokenization_utils.logging.set_verbosity_warning()

    # Log on each process the small summary:
    logger.warning(
        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
        + f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
    )

    # Detecting last checkpoint.
    last_checkpoint = None
    if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
        last_checkpoint = get_last_checkpoint(training_args.output_dir)
        if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
            raise ValueError(
                f"Output directory ({training_args.output_dir}) already exists and is not empty. "
                "Use --overwrite_output_dir to overcome."
            )
        elif last_checkpoint is not None and training_args.resume_from_checkpoint is None:
            logger.info(
                f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
                "the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
            )

    # Set seed before initializing model.
    set_seed(training_args.seed)

    config_kwargs = {
        "cache_dir": model_args.cache_dir,
        "revision": model_args.model_revision,
        "use_auth_token": True if model_args.use_auth_token else None,
    }
    if model_args.config_name:
        config = AutoConfig.from_pretrained(model_args.config_name, **config_kwargs)
    elif model_args.model_name_or_path:
        config = AutoConfig.from_pretrained(model_args.model_name_or_path, **config_kwargs)
    else:
        config = CONFIG_MAPPING[model_args.model_type]()
        logger.warning("You are instantiating a new config instance from scratch.")
        if model_args.config_overrides is not None:
            logger.info(f"Overriding config: {model_args.config_overrides}")
            config.update_from_string(model_args.config_overrides)
            logger.info(f"New config: {config}")

    tokenizer_kwargs = {
        "cache_dir": model_args.cache_dir,
        "use_fast": model_args.use_fast_tokenizer,
        "revision": model_args.model_revision,
        "use_auth_token": True if model_args.use_auth_token else None,
        "legacy": False
    }
    if model_args.tokenizer_name:
        tokenizer = AutoTokenizer.from_pretrained(model_args.tokenizer_name, **tokenizer_kwargs)
    elif model_args.tokenizer_name_or_path:
        tokenizer = AutoTokenizer.from_pretrained(model_args.tokenizer_name_or_path, **tokenizer_kwargs)
    else:
        raise ValueError(
            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
            "You can do it from another script, save it, and load it from here, using --tokenizer_name."
        )
    tokenizer.add_eos_token = True
    tokenizer.padding_side = "right"

    # Preprocessing the datasets.
    # First we tokenize all the texts.
    # since this will be pickled to avoid _LazyModule error in Hasher force logger loading before tokenize_function
    tok_logger = transformers.utils.logging.get_logger("transformers.tokenization_utils_base")

    def tokenize_function(examples):
        with CaptureLogger(tok_logger) as cl:
            output = tokenizer(examples["text"])
        # clm input could be much much longer than block_size
        if "Token indices sequence length is longer than the" in cl.out:
            tok_logger.warning(
                "^^^^^^^^^^^^^^^^ Please ignore the warning above - this long input will be chunked into smaller bits"
                " before being passed to the model."
            )
        return output
    if data_args.block_size is None:
        block_size = tokenizer.model_max_length
        if block_size > 1024:
            logger.warning(
                "The chosen tokenizer supports a `model_max_length` that is longer than the default `block_size` value"
                " of 1024. If you would like to use a longer `block_size` up to `tokenizer.model_max_length` you can"
                " override this default with `--block_size xxx`."
            )
            block_size = 1024
    else:
        if data_args.block_size > tokenizer.model_max_length:
            logger.warning(
                f"The block_size passed ({data_args.block_size}) is larger than the maximum length for the model"
                f"({tokenizer.model_max_length}). Using block_size={tokenizer.model_max_length}."
            )
        block_size = min(data_args.block_size, tokenizer.model_max_length)

    # Main data processing function that will concatenate all texts from our dataset and generate chunks of block_size.
    def group_texts(examples):
        # Concatenate all texts.
        concatenated_examples = {k: list(chain(*examples[k])) for k in examples.keys()}
        total_length = len(concatenated_examples[list(examples.keys())[0]])
        # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
        # customize this part to your needs.
        if total_length >= block_size:
            total_length = (total_length // block_size) * block_size
        # Split by chunks of max_len.
        result = {
            k: [t[i : i + block_size] for i in range(0, total_length, block_size)]
            for k, t in concatenated_examples.items()
        }
        result["labels"] = result["input_ids"].copy()
        return result
    with training_args.main_process_first(desc="dataset map tokenization and grouping"):
        lm_datasets = []
        path = Path(data_args.dataset_dir)
        files = [file.name for file in path.glob("*.txt")]
        if training_args.debug_mode is True:
            files = [files[0]]
        flag = True
        for idx, file in enumerate(files):
            data_file = os.path.join(path, file)
            filename = ''.join(file.split(".")[:-1])
            cache_path = os.path.join(data_args.data_cache_dir, filename)
            os.makedirs(cache_path, exist_ok=True)
            try:
                expect_files = []                
                if filename in expect_files:
                    logger.info(f'training datasets-{filename} has been trained')
                    continue
                else:
                    processed_dataset = datasets.load_from_disk(cache_path, keep_in_memory=False)
                    logger.info(f'training datasets-{filename} has been loaded from disk')
            except Exception:
                cache_dir = os.path.join(data_args.data_cache_dir, filename+"_text")
                os.makedirs(cache_dir, exist_ok=True)
                raw_dataset = load_dataset("text", data_files=data_file, cache_dir=cache_dir, keep_in_memory=False)
                logger.info(f"{file} has been loaded")
                tokenized_dataset = raw_dataset.map(
                    tokenize_function,
                    batched=True,
                    num_proc=data_args.preprocessing_num_workers,
                    remove_columns="text",
                    load_from_cache_file=True,
                    keep_in_memory=False,
                    cache_file_names = {k: os.path.join(cache_dir, 'tokenized.arrow') for k in raw_dataset},
                    desc="Running tokenizer on dataset",
                )
                grouped_datasets = tokenized_dataset.map(
                    group_texts,
                    batched=True,
                    num_proc=data_args.preprocessing_num_workers,
                    load_from_cache_file=True,
                    keep_in_memory=False,
                    cache_file_names = {k: os.path.join(cache_dir, 'grouped.arrow') for k in tokenized_dataset},
                    desc=f"Grouping texts in chunks of {block_size}",
                )
                processed_dataset = grouped_datasets
                processed_dataset.save_to_disk(cache_path)
            if idx == 0 or flag:
                lm_datasets = processed_dataset['train']
                flag = False
            else:
                assert lm_datasets.features.type == processed_dataset["train"].features.type
                lm_datasets = concatenate_datasets([lm_datasets, processed_dataset["train"]])

        lm_datasets = lm_datasets.train_test_split(test_size = data_args.validation_split_percentage)

    if training_args.do_train:
        train_dataset = lm_datasets['train']
        if data_args.max_train_samples is not None:
            max_train_samples = min(len(train_dataset), data_args.max_train_samples)
            train_dataset = train_dataset.select(range(max_train_samples))
        logger.info(f"Num train_samples  {len(train_dataset)}")
        logger.info("Training example:")
        logger.info(tokenizer.decode(train_dataset[0]['input_ids']))
    if training_args.do_eval:
        eval_dataset = lm_datasets["test"]
        if data_args.max_eval_samples is not None:
            max_eval_samples = min(len(eval_dataset), data_args.max_eval_samples)
            eval_dataset = eval_dataset.select(range(max_eval_samples))
        logger.info(f"Num eval_samples  {len(eval_dataset)}")
        logger.info("Evaluation example:")
        logger.info(tokenizer.decode(eval_dataset[0]['input_ids']))

    if dist.is_available() and dist.is_initialized():
        rank = torch.distributed.get_rank() % torch.cuda.device_count()
        print("GPU Rank: ", str(rank))
        os.environ["CUDA_VISIBLE_DEVICES"] = str(rank)
        device_map = {'': rank}
    else:
        device_map={'': 0}  

    if model_args.model_name_or_path:
        torch_dtype = (
            model_args.torch_dtype
            if model_args.torch_dtype in ["auto", None]
            else getattr(torch, model_args.torch_dtype)
        )
        model = AutoModelForCausalLM.from_pretrained(
            model_args.model_name_or_path,
            from_tf=bool(".ckpt" in model_args.model_name_or_path),
            config=config,
            cache_dir=model_args.cache_dir,
            revision=model_args.model_revision,
            use_auth_token=True if model_args.use_auth_token else None,
            torch_dtype=torch_dtype,
            # low_cpu_mem_usage=True,
            # load_in_8bit=False,
            # device_map=device_map
        )
    else:
        model = AutoModelForCausalLM.from_config(config)
        n_params = sum({p.data_ptr(): p.numel() for p in model.parameters()}.values())
        logger.info(f"Training new model from scratch - Total size={n_params/2**20:.2f}M params")

    model_vocab_size = model.get_output_embeddings().weight.size(0)
    tokenizer_vocab_size = len(tokenizer)
    logger.info(f"Model vocab size: {model_vocab_size}")
    logger.info(f"Tokenizer vocab size: {tokenizer_vocab_size}")
    if tokenizer_vocab_size != 55296:
        raise ValueError(f"The vocab size of tokenizer is {tokenizer_vocab_size}, not 55296. Please use Chinese-LLaMA-2 tokenizer.")
    if model_vocab_size != tokenizer_vocab_size:
        logger.info(f"Rezize model vocab size to {tokenizer_vocab_size}")
        model.resize_token_embeddings(len(tokenizer))

    # Initialize our Trainer

    def print_trainable_parameters(model, bits, detail=False, save_to=None):
        print_function = print
        """
        Prints the number of trainable parameters in the model.
        """
        trainable_params = 0
        non_trainable_params = 0
        all_params = 0
        for name, param in model.named_parameters():
            if detail:
                if "lora" not in name.lower() and param.requires_grad == True:
                    print_function(f"**PAY ATTENTION: {name}\tnumel: {param.numel()}\ttrainable:{param.requires_grad}")
                else:
                    print_function(f"{name}\tnumel: {param.numel()}\ttrainable:{param.requires_grad}\tshape: {param.shape}")
                if save_to is not None:
                    print_function(f"{name}\tnumel: {param.numel()}\ttrainable:{param.requires_grad}", file=open(save_to, 'a'))

            num_elements = param.numel() * 2 if bits == 4 else param.numel()
            all_params += num_elements
            if param.requires_grad:
                trainable_params += num_elements
            else:
                non_trainable_params += num_elements
        kbit_non_trainable_params = non_trainable_params * 2
        if bits == 4: 
            kbit_non_trainable_params = non_trainable_params / 2
        elif bits == 8:
            kbit_non_trainable_params = non_trainable_params
        print_function(
            f"========================Parameters========================\n"
            f"trainable params: {trainable_params / 1e9}B, memory requires {trainable_params * 2 / 1e9}G\n"
            f"non-trainable params: {non_trainable_params / 1e9}B, memory requeires: {kbit_non_trainable_params / 1e9}G\n"
            f"all params: {all_params / 1e9}B, memory requeires {trainable_params * 2 / 1e9 + kbit_non_trainable_params / 1e9}G\n"
            f"trainable ratio: {100.0 * trainable_params / all_params}%\n"
            f"=========================================================="
        )
        if save_to is not None:
            print_function(
            f"========================Parameters========================\n"
            f"trainable params: {trainable_params / 1e9}B, memory requires {trainable_params * 2 / 1e9}G\n"
            f"non-trainable params: {non_trainable_params / 1e9}B, memory requeires: {kbit_non_trainable_params / 1e9}G\n"
            f"all params: {all_params / 1e9}B, memory requeires {trainable_params * 2 / 1e9 + kbit_non_trainable_params / 1e9}G\n"
            f"trainable ratio: {100.0 * trainable_params / all_params}%\n"
            f"==========================================================", file=open(save_to, 'a')
        )

    print_trainable_parameters(model, 16)

    trainer = Trainer(
        model=model,
        args=training_args,
        train_dataset=train_dataset if training_args.do_train else None,
        eval_dataset=eval_dataset if training_args.do_eval else None,
        tokenizer=tokenizer,
        data_collator=fault_tolerance_data_collator,
        compute_metrics=compute_metrics if training_args.do_eval and not is_torch_tpu_available() else None,
        preprocess_logits_for_metrics=preprocess_logits_for_metrics
        if training_args.do_eval and not is_torch_tpu_available()
        else None,
    )
    trainer.add_callback(SaveModelCallback)
    # Training

    if training_args.do_train:
        checkpoint = None
        if training_args.resume_from_checkpoint is not None:
            checkpoint = training_args.resume_from_checkpoint
        elif last_checkpoint is not None:
            checkpoint = last_checkpoint
        with torch.autocast("cuda"):
            train_result = trainer.train(resume_from_checkpoint=checkpoint)
        trainer.save_model(os.path.join(training_args.output_dir, "pt_model"))

        metrics = train_result.metrics

        max_train_samples = (
            data_args.max_train_samples if data_args.max_train_samples is not None else len(train_dataset)
        )
        metrics["train_samples"] = min(max_train_samples, len(train_dataset))

        trainer.log_metrics("train", metrics)
        trainer.save_metrics("train", metrics)
        trainer.save_state()

    # Evaluation
    if training_args.do_eval:
        logger.info("*** Evaluate ***")

        with torch.autocast("cuda"):
            metrics = trainer.evaluate()

        max_eval_samples = data_args.max_eval_samples if data_args.max_eval_samples is not None else len(eval_dataset)
        metrics["eval_samples"] = min(max_eval_samples, len(eval_dataset))
        try:
            perplexity = math.exp(metrics["eval_loss"])
        except OverflowError:
            perplexity = float("inf")
        metrics["perplexity"] = perplexity

        trainer.log_metrics("eval", metrics)
        trainer.save_metrics("eval", metrics)

if __name__ == "__main__":
    main()

提交任务脚本：

lr=2e-4
pretrained_model=/code/xx/LLM_mine/model/LLama2/Llama2_7b/llama2_7b
chinese_tokenizer_path=/code/xx/LLM_mine/reference/Chinese-LLaMA-Alpaca-2/scripts/tokenizer
dataset_dir=/code/xx/LLM_mine/data/pretrain_test
data_cache=/code/xx/LLM_mine/data/pt_zh_en_txt_data_cache
per_device_train_batch_size=1
per_device_eval_batch_size=1
gradient_accumulation_steps=8
output_dir=/code/xx/LLM_mine/scripts/llama2_pretrain/output_lora/lora_full_fp16

deepspeed_config_file=/code/xx/LLM_mine/scripts/llama2_pretrain/ds_zero3_no_offload.json
MASTER_ADDR=$(ping -c 1 ${MASTER_ADDR} |grep PING | grep -E -o '([0-9]{1,3}\.){3}[0-9]{1,3}')

torchrun --nnodes 1 --nproc_per_node 2 --rdzv_id=100 --rdzv_backend=c10d --rdzv_endpoint=${MASTER_ADDR}:29400 /code/xx/LLM_mine/scripts/llama2_pretrain/run_clm_pt_with_full.py \
    --deepspeed ${deepspeed_config_file} \
    --model_name_or_path ${pretrained_model} \
    --tokenizer_name_or_path ${chinese_tokenizer_path} \
    --dataset_dir ${dataset_dir} \
    --data_cache_dir ${data_cache} \
    --validation_split_percentage 0.001 \
    --per_device_train_batch_size ${per_device_train_batch_size} \
    --per_device_eval_batch_size ${per_device_eval_batch_size} \
    --do_train \
    --seed 666 \
    --fp16 \
    --bf16 False \
    --num_train_epochs 1 \
    --lr_scheduler_type cosine \
    --learning_rate ${lr} \
    --warmup_ratio 0.05 \
    --weight_decay 0.005 \
    --logging_strategy steps \
    --logging_steps 10 \
    --save_strategy steps \
    --save_total_limit 2 \
    --save_steps 20 \
    --gradient_accumulation_steps ${gradient_accumulation_steps} \
    --preprocessing_num_workers 8 \
    --block_size 1024 \
    --output_dir ${output_dir} \
    --overwrite_output_dir \
    --ddp_timeout 30000 \
    --logging_first_step True \
    --torch_dtype float16 \
    --gradient_checkpointing \
    --ddp_find_unused_parameters False \
    --flash_attn False \

依赖情况（代码类问题务必提供）

 RUN pip install git+https://github.com/huggingface/peft.git@13e53fc
 RUN pip install transformers==4.31.0
 RUN pip install sentencepiece==0.1.97
 RUN pip install bitsandbytes==0.39.1
 RUN pip install xformers
 RUN MAX_JOBS=2 pip install flash-attn --no-build-isolation -i https://mirrors.aliyun.com/pypi/simple

运行日志或截图

保存下来的bin文件参数都是空

[airaria]

https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/blob/b1c9d8cf09e614fe37f374b4e9fd4eaa31545fbe/scripts/training/run_clm_pt_with_peft.py#L587-L590

这部分代码注释了吗

[Double-bear]

https://github.com/ymcui/Chinese-LLaMA-Alpaca-2/blob/b1c9d8cf09e614fe37f374b4e9fd4eaa31545fbe/scripts/training/run_clm_pt_with_peft.py#L587-L590

这部分代码注释了吗

注释了，代码里没有这部分了

[TengLi931128]

trainer.add_callback(SaveModelCallback) 这个是也注释掉试试？

[Double-bear]

trainer.add_callback(SaveModelCallback) 这个是也注释掉试试？

因为我需要将checkpoint存成huggingface格式的模型，所以我加上了这个callback函数，但是正常存zero的pth格式的文件是有的

[Double-bear]

我似乎找到问题在哪里，因为我是使用zero3进行训练，我发现模型的vacab size有问题：

• INFO - main - Model vocab size: 0
• INFO - main - Tokenizer vocab size: 55296
• INFO - main - Rezize model vocab size to 55296 当我换成zero2的时候，就不会出现：
• INFO - main - Model vocab size: 32000
• INFO - main - Tokenizer vocab size: 55296

• INFO - main - Rezize model vocab size to 55296 请问有大佬知道这个是什么原因吗？

  
  # 这是zero3配置
  {
  "fp16": {
      "enabled": "auto",
      "loss_scale": 0,
      "loss_scale_window": 100,
      "initial_scale_power": 16,
      "hysteresis": 2,
      "min_loss_scale": 1e-10
  },
  
  "bf16": {
      "enabled": "auto"
  },
  
  "zero_optimization": {
      "stage": 3,
      "allgather_partitions": true,
      "allgather_bucket_size": 1e8,
      "overlap_comm": true,
      "reduce_scatter": true,
      "reduce_bucket_size": 1e8,
      "contiguous_gradients": true
  
  },
  
  "gradient_accumulation_steps": "auto",
  "gradient_clipping": "auto",
  "steps_per_print": 2000,
  "train_batch_size": "auto",
  "train_micro_batch_size_per_gpu": "auto",
  "wall_clock_breakdown": false
  }

这是zero2配置

{ "fp16": { "enabled": "auto", "loss_scale": 0, "loss_scale_window": 100, "initial_scale_power": 16, "hysteresis": 2, "min_loss_scale": 1e-10 },

"bf16": {
    "enabled": "auto"
},

"zero_optimization": {
    "stage": 2,
    "allgather_partitions": true,
    "allgather_bucket_size": 1e8,
    "overlap_comm": true,
    "reduce_scatter": true,
    "reduce_bucket_size": 1e8,
    "contiguous_gradients": true

},

"gradient_accumulation_steps": "auto",
"gradient_clipping": "auto",
"steps_per_print": 2000,
"train_batch_size": "auto",
"train_micro_batch_size_per_gpu": "auto",
"wall_clock_breakdown": false

}

[TengLi931128]

我的经验是，即使不加trainer.add_callback(SaveModelCallback)，trainer保存的checkpoint也是.bin格式的，我不知道你为什么会保存成.pth格式。 另外，zero3和lora貌似不能同时使用。但按理来说，全参数是不用lora的，你不应该遇到这个问题才对。

[Double-bear]

我的经验是，即使不加trainer.add_callback(SaveModelCallback)，trainer保存的checkpoint也是.bin格式的，我不知道你为什么会保存成.pth格式。 另外，zero3和lora貌似不能同时使用。但按理来说，全参数是不用lora的，你不应该遇到这个问题才对。

我用deepspeed训练checkpoint保存的就是rngstate*.pth格式的文件，这个好像是因为deepspeed的原因。我没有用zero3训练lora，我训的是全部参数，而且使用zero3训练读取模型后model vocab size就是0，就很奇怪。

[iMountTai]

@Double-bear 在你的zero3 config 里面设置stage3_gather_16bit_weights_on_model_save为true就可以正常保存pytorch_model.bin了。使用zero3训练读取模型后model vocab size为0是正常的，不耽误训练即可

[Double-bear]

stage3_gather_16bit_weights_on_model_save

非常感谢，已经可以正常保存了。

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your consideration.