open_dutch_llm_leaderboard / evals /hellaswag /hellaswag_nl_phi-2.json
Bram Vanroy
add evals
d20fb33
raw
history blame
558 Bytes
{
"results": {
"hellaswag_nl": {
"acc": 0.2746896923907178,
"acc_stderr": 0.004637499043161815,
"acc_norm": 0.3019967620075553,
"acc_norm_stderr": 0.004770134104673541
}
},
"versions": {
"hellaswag_nl": 1
},
"config": {
"model": "hf-auto",
"model_args": "pretrained=microsoft/phi-2,use_accelerate=True,device_map_option=auto,dtype=auto,load_in_8bit=True",
"batch_size": 4,
"device": "cuda",
"no_cache": false,
"limit": null,
"bootstrap_iters": 100000,
"description_dict": {}
}
}