Views
No views yet
onnx/model_quantized.onnx — int8 quantized weights (preferred for browser)config.json / ort_config.json / inference_config.json1{
2 "threshold": 0.55,
3 "base_model": "allenai/scibert_scivocab_uncased",
4 "max_len": 96,
5 "test_precision": 0.9414289028067734,
6 "test_recall": 0.8224744148343297,
7 "test_f1": 0.8779406197609648,
8 "test_roc_auc": 0.9571438004475947,
9 "n_train": 48000
10}