Zero-shot Evaluation
Duncan1115 opened this issue · comments
Juncan Deng commented
Hi, thanks for your great work!
I wonder how to evaluate the pruned and fintuned model in lm-evaluation-harness?
Horseee commented
Hi.
We release the code, and you can now follow the instructions to run the evaluation😄
We haven't comprehensively tested whether the current evaluation code works well on different devices and on different version of PyTorch, so if you find any problem, feel free to ask me!
Juncan Deng commented
Thanks! It helps me a lot.