New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[问题反馈]SFT后的checkpoint加载目录 #249

Closed

hunter-xue opened this issue Dec 25, 2023 · 1 comment

hunter-xue commented Dec 25, 2023 •

edited

Loading

做Qwen-7B的SFT后，输出目录如下：

然后通过swift app-ui --ckpt_dir进行推理测试：

如果ckpt_dir设置为/data/Qwen/output-Qwen/qwen-7b-chat/v4-20231224-173846，可以正常启动，但模型使用的仍然是原始模型
如果ckpt_dir设置为/data/Qwen/output-Qwen/qwen-7b-chat/v4-20231224-173846/checkpoint-100，则可以正常加载SFT之后的checkpoint

可否在启动过程打印一个提示，告知加载的是原始模型还是SFT之后的checkpoint. 或者在文档中说明一下加载哪个目录是正确的。

The text was updated successfully, but these errors were encountered:

Collaborator

tastelikefeet commented Dec 25, 2023

好的，这个我添加一下提示防止出错

tastelikefeet added a commit to tastelikefeet/swift that referenced this issue


          fix issue modelscope#249

377beef

tastelikefeet added a commit that referenced this issue


          fix issue #249 (#250)

a1abff5

tastelikefeet closed this as completed

tastelikefeet added a commit to tastelikefeet/swift that referenced this issue


          Merge commit 'f3e3631fc520d0d48853539e52e586e514a5437f' into feat/age…

e257bba

…nt_and_rl

* commit 'f3e3631fc520d0d48853539e52e586e514a5437f':
  Update readme for SCEdit (modelscope#258)
  fix unicode error (modelscope#259)
  Update 1228 (modelscope#254)
  Feat/scedit (modelscope#253)
  fix issue modelscope#249 (modelscope#250)
  Add sft for codegeex2 (modelscope#248)
  Fix copying additional files (modelscope#247)
  Support more peft tuners (modelscope#245)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment