This model is a fork of bert-large-uncased-whole-word-masking-finetuned-squad. ONNX and OpenVINO-IR models are enclosed.

evaluated in v4.9.2.

  eval_exact_match = 86.9253
  eval_f1          = 93.1563
  eval_samples     =   10784

Dataset used to train vuiseng9/bert-l-squadv1.1-sl384

Evaluation results