You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
为什么我微调Qwen2-7B-Instruct,说没有这个id呢?
File "/home/powerop/work/conda/envs/swift/lib/python3.10/site-packages/swift/llm/utils/argument.py", line 762, in set_model_type
raise ValueError(f"model_type: '{args.model_type}' is not registered. "
为什么我微调Qwen2-7B-Instruct,说没有这个id呢? File "/home/powerop/work/conda/envs/swift/lib/python3.10/site-packages/swift/llm/utils/argument.py", line 762, in set_model_type raise ValueError(f"model_type: '{args.model_type}' is not registered. "
使用swift对Qwen2-72B-Chat进行自我认知微调,让模型认为自己是小胡,由魔搭训练。
在开始微调之前,需要进行环境准备:
我们使用swift提供的带模型名字和作者通配符的self-cognition数据集进行自我认知微调,以及使用alpaca-zh、alpaca-en数据集保持模型的通用能力。整个微调过程大约需要30分钟,微调脚本如下:
微调的超参数含义可以参考命令行参数文档:https://github.com/modelscope/swift/blob/main/docs/source/LLM/%E5%91%BD%E4%BB%A4%E8%A1%8C%E5%8F%82%E6%95%B0.md
微调过程的loss可视化:
![image](https://private-user-images.githubusercontent.com/45290347/337343928-c2f38585-08e4-4879-bbf5-3da2785d8eb0.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.gfTe4rz2P3sfH3j0IxBqsLDBBKWTsx4pEMrrEYWo_Pw)
微调显存占用:
![image](https://private-user-images.githubusercontent.com/45290347/337343974-3a620930-e89f-4a16-9314-b1fd32f99c87.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.IBezJOpSNP7yur3XvLAoPO5hMcJBGTdfOETSHF-A1CI)
微调后推理脚本如下,这里的ckpt_dir需要修改为微调生成的checkpoint文件夹:
训练后效果:
![image](https://private-user-images.githubusercontent.com/45290347/337344068-0b6ba272-fd93-406b-bce1-dab787f1c545.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.4-6Cinqvji5INLyIYVTvmMhtO0n8AfYL5efVEC4Z474)
The text was updated successfully, but these errors were encountered: