I expect the TrainingArguments class from the Hugging Face transformers library to initialize without any ImportError when I have installed all required dependencies (transformers, accelerate, and PyTorch). Specifically, after running the following commands:
System Info
transformers
version: 4.44.2Who can help?
No response
Information
Tasks
examples
folder (such as GLUE/SQuAD, ...)Reproduction
from transformers import TrainingArguments training_args = TrainingArguments( output_dir='./output', do_train=True, do_eval=True, num_train_epochs=100,
per_device_train_batch_size=32,
per_device_eval_batch_size=16, warmup_steps=100,
weight_decay=0.05, logging_strategy='steps', logging_dir='./multi-class-logs',
logging_steps=50, evaluation_strategy="steps", eval_steps=50, save_strategy="steps", load_best_model_at_end=True )
Expected behavior
Expected Behavior:
I expect the TrainingArguments class from the Hugging Face transformers library to initialize without any ImportError when I have installed all required dependencies (transformers, accelerate, and PyTorch). Specifically, after running the following commands:
pip install transformers[torch] pip install accelerate -U