NameError: name 'PartialState' is not defined error while training hugging face wave2vec model

以下是导致错误的代码块

training_args = TrainingArguments(
    output_dir="my_awesome_mind_model",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=3e-5,
    per_device_train_batch_size=32,
    gradient_accumulation_steps=4,
    per_device_eval_batch_size=32,
    num_train_epochs=10,
    warmup_ratio=0.1,
    logging_steps=10,
    load_best_model_at_end=True,
    metric_for_best_model="accuracy",
    push_to_hub=True,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=dataset["train"],
    # eval_dataset=encoded_minds["test"],
    tokenizer=feature_extractor,
    compute_metrics=compute_metrics,
)

trainer.train()

遇到以下错误

NameError Traceback (most recent call last) in <cell line: 1>() 1 training_args = TrainingArguments( 2 output_dir="my_awesome_mind_model", 3 evaluation_strategy="epoch", 4 save_strategy="epoch", 5 learning_rate=3e-5,

4 frames /usr/local/lib/python3.10/dist-packages/transformers/training_args.py in _setup_devices(self) 1629 self._n_gpu = 1 1630 else: 1631 self.distributed_state = PartialState(backend=self.ddp_backend) 1632 self._n_gpu = 1 1633 if not is_sagemaker_mp_enabled():

NameError: name 'PartialState' is not defined

我试图在另一个数据集上遵循拥抱脸部的音频分类指南（链接），但在运行训练args代码时，我得到了名称"PartialState"未定义的错误。

截止到2023年5月11日：

这个错误似乎是由huggingface/accelerate库中的仍未解决的问题引起的。

你可以尝试以下的解决方案：

重新安装transformers & accelerate

pip uninstall -y transformers accelerate
pip install transformers accelerate

如果你使用colab/Jupyter，确保重新启动笔记本的Runtime

安装accelerate的开发版本。

pip install git+https://github.com/huggingface/accelerate

倒退到以前的transformers 版本（4.28.0）

# You might also need to uninstall transformers first: pip uninstall -y transformers
pip install transformers==4.28.0

pip install --upgrade accelerate并重新启动你的笔记本的runtime。