[BUG/Help] <pytorch_model.bin don't exists>

[sc-carson]

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

使用ptuning进行微调后，在checkpoint-3000路径下并没有pytorch_model.bin，只有training_args.bin，无法运行prefix_state_dict = torch.load(os.path.join(CHECKPOINT_PATH, "pytorch_model.bin"))

Expected Behavior

No response

Steps To Reproduce

bash train.sh

Environment

- OS:Ubuntu 22.04
- Python:Python 3.10.13
- Transformers:
- PyTorch:4.36.2
- CUDA Support (`python -c "import torch; print(torch.cuda.is_available())"`) :True

Anything else?

No response

[bazinga014]

我也是，请问解决了吗

[liumazeze]

同，请问有无解决方案？

[liumazeze]

已解决 -> transforms库降到 requirements.txt 中的版本即可解决