Hugging Face's logo

hung200504
/
bert-covidqa-5

Question Answering
Transformers
PyTorch
deberta-v2
Generated from Trainer
Inference Endpoints
Model card Files Community
bert-covidqa-5 / README.md
hung200504's picture
hung200504
bert-cased
58fb031
preview code
|
raw
history blame contribute delete
2.43 kB
metadata
license: mit
base_model: timpal0l/mdeberta-v3-base-squad2
tags:
  - generated_from_trainer
datasets:
  - covid_qa_deepset
model-index:
  - name: bert-covidqa-5
    results: []

bert-covidqa-5

This model is a fine-tuned version of timpal0l/mdeberta-v3-base-squad2 on the covid_qa_deepset dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4190

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 3e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
0.5956 0.04 5 0.4016
0.3741 0.09 10 0.3879
0.3405 0.13 15 0.4240
0.4372 0.18 20 0.4102
0.2592 0.22 25 0.4534
0.3534 0.26 30 0.4571
0.4268 0.31 35 0.4107
0.2663 0.35 40 0.4166
0.143 0.39 45 0.4345
0.2494 0.44 50 0.5575
0.8953 0.48 55 0.6172
0.5504 0.53 60 0.4879
0.6411 0.57 65 0.3718
0.5454 0.61 70 0.3929
0.4441 0.66 75 0.3641
0.2922 0.7 80 0.3638
0.491 0.75 85 0.3785
0.4362 0.79 90 0.3938
0.1633 0.83 95 0.4162
0.6762 0.88 100 0.4321
0.3111 0.92 105 0.4241
0.3453 0.96 110 0.4190

Framework versions

  • Transformers 4.34.1
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.5
  • Tokenizers 0.14.1