open_dutch_llm_leaderboard / evals /hellaswag /hellaswag_nl_Orca-2-7b.json
Bram Vanroy
add evaluations
1303fdd
raw
history blame
571 Bytes
{
"results": {
"hellaswag_nl": {
"acc": 0.39071775499190503,
"acc_stderr": 0.00506922492446768,
"acc_norm": 0.4887209929843497,
"acc_norm_stderr": 0.005193500820957502
}
},
"versions": {
"hellaswag_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=microsoft/Orca-2-7b,use_accelerate=True,device_map_option=auto,dtype=bfloat16,load_in_8bit=True",
"batch_size": "auto",
"device": "cuda",
"no_cache": false,
"limit": null,
"bootstrap_iters": 100000,
"description_dict": {}
}
}