BAOOOOOM

Results 7 comments of BAOOOOOM

I find three part is indispensability: 'prompt_model.template.wte.weight', 'prompt_model.template.control_trans.0.weight','prompt_model.template.control_trans.2.weight'. Without anyone, the prefix-tuning model will arise the performance like untrained model. But I still not find which part is add to...

I used the prefix-t5-3b model to classify.

I haven't used multi-GPU techniques. This is the code, thanks: ` import argparse import torch parser = argparse.ArgumentParser("") parser.add_argument("--lr", type=float, default=1e-5) parser.add_argument("--plm_eval_mode", action="store_true") parser.add_argument("--model", type=str, default='t5') # tested model are...

And the environment is python3.8, torch==1.11.0, transformers==4.20.1, openprompt==1.0.0

> Sorry, I haven't found a solution for that.

![image](https://user-images.githubusercontent.com/82091145/197475657-d25ea6da-bfc5-4840-8fb1-cd14ae35717e.png)