Skip to content
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.

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

3版本更新后 训练model type使用方式发生变化 #3108

Open
EvilCalf opened this issue Feb 14, 2025 · 4 comments
Open

3版本更新后 训练model type使用方式发生变化 #3108

EvilCalf opened this issue Feb 14, 2025 · 4 comments

Comments

@EvilCalf
Copy link

请问下,现在3.1,我要从qwen2.5-base上全参训sft,用的config是base,这个对最终对训练有影响,stop token的判定也有问题。训练出来最终的配置里eos token有问题

@Jintao-Huang
Copy link
Collaborator

使用swift infer推理正常嘛

@Jintao-Huang
Copy link
Collaborator

qwen2.5-base上全参训sft,默认也会使用qwen2.5的template。里面有设置stop_words,只是不对config.json进行修改

@EvilCalf
Copy link
Author

推理用vllm和sglang都试过,不加stop=["<|im_end|>", "<|endoftext|>"],无法正常停止,看日志已经预测出了停止token,但依旧在进行预测,但无法看到后续的token

@EvilCalf
Copy link
Author

这个问题在之前2.x版本没有遇到过,是正常用base 基于instruct的config进行训练的

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants