belle9217
belle9217
我用base v100 32G都能OOM
对啊,base v100 32G 都能OOM
the same problem I encounter! how do you resolved?
我也遇到这个问题,请问你的问题解决了吗
how to resolve the bug?thanks for your work.
i got the same error, do you resolve it ?
yes,i have the same error use the command: m_eval --model hf --model_args pretrained=dc3.5-7b-32k --tasks mmlu --device cuda:4 --batch_size 8 --num_fewshot 5 --output_path ./eval_out --use_cache ./eval_cache_mmlu
@dz1iang yes,it works,that is all I need.
> qwen2 huggingface model 评测结果有问题  我得到的也是这个结果,请问您解决了吗