I followed your guideline to finetune the model and try it. But the saved models were only these stuff (image below):
So, when I loaded the PEFT model to use, it could not find the config.json of the model. How can I get over it?
Note that this is the checkpoint saved by Huggingface Trainer model during training. I canceled the training job so .save_pretrained() method in the end might not execute.
Thanks for your interesting work!
I followed your guideline to finetune the model and try it. But the saved models were only these stuff (image below):
So, when I loaded the PEFT model to use, it could not find the config.json of the model. How can I get over it?
Note that this is the checkpoint saved by Huggingface Trainer model during training. I canceled the training job so .save_pretrained() method in the end might not execute.