multi-motions-28 / benchmarks.json
proxy3d's picture
Upload folder using huggingface_hub
299bfda verified
Raw
History Blame Contribute Delete
1.92 kB
{
"author": {
"name": "Ilya Zelenskiy",
"alias": "proxy3d",
"telegram": "https://t.me/greenruff"
},
"project_links": {
"communication_styles_article": "https://iproxy3d.github.io/communication-styles-llm/",
"communication_styles_github": "https://github.com/iproxy3d/communication-styles-llm"
},
"evaluation_protocol": "3 independent seeds; native-RU benchmarks were not used for training, checkpoint selection, or threshold tuning",
"goemotions_test": {
"en": {
"micro_f1_mean": 0.595278,
"micro_f1_std": 0.001725,
"macro_f1_mean": 0.519635,
"macro_f1_std": 0.00347,
"micro_ap_mean": 0.63659,
"micro_ap_std": 0.00412,
"macro_ap_mean": 0.506498,
"macro_ap_std": 0.002801
},
"ru_translated": {
"micro_f1_mean": 0.565428,
"micro_f1_std": 0.001362,
"macro_f1_mean": 0.482554,
"macro_f1_std": 0.005028,
"micro_ap_mean": 0.594971,
"micro_ap_std": 0.002251,
"macro_ap_mean": 0.46094,
"macro_ap_std": 0.005664
}
},
"native_ru": {
"cedr": {
"n": 1882,
"labels": [
"joy",
"sadness",
"surprise",
"fear",
"anger"
],
"micro_f1_mean": 0.517982,
"micro_f1_std": 0.013487,
"macro_f1_mean": 0.529451,
"macro_f1_std": 0.018099,
"micro_ap_mean": 0.638756,
"micro_ap_std": 0.000341,
"macro_ap_mean": 0.676202,
"macro_ap_std": 0.007899
},
"semeval2025_ru": {
"n": 1000,
"labels": [
"anger",
"disgust",
"fear",
"joy",
"sadness",
"surprise"
],
"micro_f1_mean": 0.63545,
"micro_f1_std": 0.014176,
"macro_f1_mean": 0.633188,
"macro_f1_std": 0.017979,
"micro_ap_mean": 0.840718,
"micro_ap_std": 0.001325,
"macro_ap_mean": 0.844767,
"macro_ap_std": 0.004708
}
}
}