We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
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
您好,您提供的数据集中,DISC-Law-SFT-Triplet 包含 input,output 和 reference 三个部分。在用 LLaMA Efficient Tuning 微调时,请问 reference 是如何加入训练的呢?我目前是把它作为 system 输入,或者说这部分应该直接拼接到 input 中?
input
output
reference
The text was updated successfully, but these errors were encountered:
如技术报告里说的,reference作为context 拼在input中作为模型输入
Sorry, something went wrong.
明白了,谢谢。 顺便想请教一下,您README中给出的LoRA微调的指令,最终得到的结果如何?我在Baichuan2-7B和Qwen1.5-7B两个个模型上分别尝试了LoRA微调,学习率等超参没变,benchmark跑出来比微调前的原始模型要差不少。请问您有在这些模型上测试过吗?
No branches or pull requests
您好,您提供的数据集中,DISC-Law-SFT-Triplet 包含
input
,output
和reference
三个部分。在用 LLaMA Efficient Tuning 微调时,请问reference
是如何加入训练的呢?我目前是把它作为 system 输入,或者说这部分应该直接拼接到 input 中?The text was updated successfully, but these errors were encountered: