Something wrong in the data preprocess about chatglm3-finetune

[xxw1995]

Check the 20th line in tokenize_dataset_rows.py and see if the path to the model is correct. If the path is correct, then check if the model weights are correct.

[xxw1995]

If the above model does not have any issues, please check the version of the datasets.If there are no issues with the model mentioned above, check the version of the dataset. datasets==2.10.1

[yt7589]

Solved! That's greate. Thank you for your quick response!
I following your advise change the mode value in tokenize_dataset_rows.py. It worked. I had change the model value in finetune.py too. The finetunning process was succeed too.
This is a greate job.
Do you have plan to support QLoRA?

By the way you had better add the instruction to change the model value in tokenize_dataset_rows.py and finetune.py in the README.md.