opencompass
opencompass copied to clipboard
[Feature]在model文件中没有找到关于Qwen1.5-32b-chat的配置文件
Describe the feature
只有7b,14b,72b。这个怎么办呢?
Will you implement it?
- [ ] I would like to implement this feature and create a PR!
You can change the path with 32b-chat, and we will update the new model config soon.
好的谢谢!我自己尝试把微调后32b的qwen模型转成hf格式,然后运行命令如下python run.py --datasets cmmlu_gen
--hf-path /root/autodl-tmp/Qwen1.5-32B-Chat
--model-kwargs device_map='auto'
--tokenizer-kwargs padding_side='left' truncation='left' trust_remote_code=True
--max-seq-len 300
--max-out-len 5
--batch-size 8
--num-gpus 1
。请问我还需要添加其他参数设置吗?比如模板什么的?