train.sh: line 3: $'\r': command not found
Traceback (most recent call last):
File "/mnt/e/cqy-gpt/gpt-fine-tune/ChatGLM-6B/ptuning/main.py", line 26, in
import numpy as np
ModuleNotFoundError: No module named 'numpy'
train.sh: line 5: --do_train: command not found
train.sh: line 6: --train_file: command not found
train.sh: line 7: --validation_file: command not found
train.sh: line 8: --prompt_column: command not found
train.sh: line 9: --response_column: command not found
train.sh: line 10: --overwrite_cache: command not found
train.sh: line 11: --model_name_or_path: command not found
train.sh: line 12: --output_dir: command not found
train.sh: line 13: --overwrite_output_dir: command not found
train.sh: line 14: --max_source_length: command not found
train.sh: line 15: --max_target_length: command not found
train.sh: line 16: --per_device_train_batch_size: command not found
train.sh: line 17: --per_device_eval_batch_size: command not found
train.sh: line 18: --gradient_accumulation_steps: command not found
train.sh: line 19: --predict_with_generate: command not found
train.sh: line 20: --max_steps: command not found
train.sh: line 21: --logging_steps: command not found
train.sh: line 22: --save_steps: command not found
train.sh: line 23: --learning_rate: command not found
train.sh: line 24: --pre_seq_len: command not found
train.sh: line 25: --quantization_bit: command not found
train.sh: line 26: $'\r': command not found
- git clone https://github.com/THUDM/ChatGLM-6B.git
- 下载chatglm的模型
- 下载ADGEN数据
- cd ptuning
- bash train.sh
Environment
- OS:windows 11
- Python:3.10.11
- Transformers:4.27.1
- PyTorch:2.0.0+cpu
- CUDA Support (`python -c "import torch; print(torch.cuda.is_available())"`) :True