Closed Zkli-hub closed 4 days ago
llamafactory
CUDA_VISIBLE_DEVICES=0 llamafactory-cli eval --model_name_or_path PATH --template fewshot --task mmlu --split test --lang en --n_shot 5 --batch_size 1
Evaluate the MMLU performance of llama2-7b-base, but the MMLU results is lower than the original paper which is 45.3
No response
A small derivation is acceptable and usual
Reminder
System Info
llamafactory
version: 0.8.3.dev0Reproduction
CUDA_VISIBLE_DEVICES=0 llamafactory-cli eval --model_name_or_path PATH --template fewshot --task mmlu --split test --lang en --n_shot 5 --batch_size 1
Expected behavior
Evaluate the MMLU performance of llama2-7b-base, but the MMLU results is lower than the original paper which is 45.3![image](https://github.com/hiyouga/LLaMA-Factory/assets/55663065/0660faf4-e71b-4a40-9cf5-c42a5cf2fc50)
Others
No response