How to use it to test models?

[cyzhh]

Hello, I feel that your github is very difficult to use. The eval.py type ”bai-scieval-valid.json“ does not have this file. In addition, can you improve your README file to better test other open source models such as ChatGLM?

[Sunliangtai]

Sorry for the llllllllate reply, since this is the official github account of our lab, I didn't receive your question via email in time. Thanks a lot! I have modified eval.py and fix the bug you mentioned. To test other open source models, you can evaluate using scieval-test.json and then upload the results to our website.