Training Script For Step 3 #34
Description
Hello,
Thank you so much for sharing such great work!
I'm currently on the step 3, which is time to train the pre-experienced model. I found there is a train.py from training/stanford_alpaca in your repo. So, is this your training script? Also, according to your implementation section from your paper, you said your training was based on the FastChat codebase. And I found there is only full-parameter fine-tuning from that codebase. So, am I correct in understanding that the Step 3 essentially corresponds to full-parameter fine-tuning?
Looking forward to your response!
Cheers!