opencompass
opencompass copied to clipboard
[Feature] ModuleNotFoundError: No module named 'opencompass.datasets.lawbench.evaluation_functions'
Describe the feature
我执行命令python run.py --datasets ceval_ppl mmlu_ppl --hf-path /T106/LLM_model/llama-7b --model-kwargs device_map='auto' --tokenizer-kwargs padding_side='left' truncation='left' use_fast=False --max-out-len 100 --max-seq-len 2048 --batch-size 8 --no-batch-padding --num-gpus 1
...................................
98%|████████████████████████████████████████████████████████████████████████████████████████████▎ | 107/109 [04:13<00:01, 1.99it/s]12/25 12:59:23 - OpenCompass - WARNING - task OpenICLEval[opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/lukaemon_mmlu_conceptual_physics] fail, see
./outputs/default/20231225_125408/logs/eval/opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/lukaemon_mmlu_conceptual_physics.out
12/25 12:59:23 - OpenCompass - WARNING - task OpenICLEval[opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/lukaemon_mmlu_us_foreign_policy] fail, see
./outputs/default/20231225_125408/logs/eval/opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/lukaemon_mmlu_us_foreign_policy.out
100%|██████████████████████████████████████████████████████████████████████████████████████████████| 109/109 [04:13<00:00, 2.33s/it]
12/25 12:59:23 - OpenCompass - ERROR - /T106/model_main/opencompass-main/opencompass-main/opencompass/runners/base.py - summarize - 63 - OpenICLEval[opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/ceval-computer_network] failed with code 1
12/25 12:59:23 - OpenCompass - ERROR - /T106/model_main/opencompass-main/opencompass-main/opencompass/runners/base.py - summarize - 63 - OpenICLEval[opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/ceval-operating_system] failed with code 1
...................
...................
...................
12/25 12:59:23 - OpenCompass - ERROR - /T106/model_main/opencompass-main/opencompass-main/opencompass/runners/base.py - summarize - 63 - OpenICLEval[opencompass.models.huggingface.HuggingFace_LLM_model_llama-7b/lukaemon_mmlu_conceptual_physics] failed with code 1
dataset
...............................
File "
Will you implement it?
在跑mixtral-8x7b-v0.1的时候遇到同样的问题,想问下预计什么时候支持呢?
same