quastrinos's picture
Upload TFDebertaV2ForMultipleChoice
80add85
metadata
license: mit
base_model: microsoft/deberta-v3-large
tags:
  - generated_from_keras_callback
model-index:
  - name: TF-40k-openbook-finetuned-deberta-v3-large-mcqa-TPU-v2
    results: []

TF-40k-openbook-finetuned-deberta-v3-large-mcqa-TPU-v2

This model is a fine-tuned version of microsoft/deberta-v3-large on an unknown dataset. It achieves the following results on the evaluation set:

  • Train Loss: 0.6671
  • Validation Loss: 0.9061
  • Train Map@3: 0.8095
  • Epoch: 2

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:

  • optimizer: {'name': 'Adam', 'weight_decay': 0.01, 'clipnorm': 1, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': False, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'CosineDecay', 'config': {'initial_learning_rate': 2e-06, 'decay_steps': 2826, 'alpha': 5e-09, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
  • training_precision: mixed_bfloat16

Training results

Train Loss Validation Loss Train Map@3 Epoch
0.9561 0.8899 0.8073 0
0.7125 0.8513 0.8244 1
0.6671 0.9061 0.8095 2

Framework versions

  • Transformers 4.35.0.dev0
  • TensorFlow 2.12.0
  • Datasets 2.14.5
  • Tokenizers 0.14.1