LLM_Teached_Bart / README.md
GlycerinLOL's picture
End of training
cc6d357
|
raw
history blame
2.3 kB
metadata
license: mit
base_model: facebook/bart-large-xsum
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: LLM_Teached_Bart
    results: []

LLM_Teached_Bart

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

  • Loss: 2.3237
  • Rouge1: 0.4756
  • Rouge2: 0.203
  • Rougel: 0.3677
  • Rougelsum: 0.3678
  • Gen Len: 41.4318

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: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 8
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
1.6644 1.0 1250 1.6972 0.4687 0.2036 0.3619 0.362 43.4245
1.3035 2.0 2500 1.6463 0.4762 0.2104 0.3746 0.3747 39.5091
1.0206 3.0 3750 1.7278 0.476 0.2117 0.3743 0.3746 38.9555
0.8224 4.0 5000 1.8642 0.477 0.2094 0.3724 0.3723 40.5182
0.654 5.0 6250 1.9480 0.4757 0.2083 0.3717 0.3716 39.8736
0.5302 6.0 7500 2.1332 0.4773 0.2062 0.37 0.3699 40.8309
0.4364 7.0 8750 2.2474 0.4749 0.2008 0.3648 0.3648 42.0391
0.3782 8.0 10000 2.3237 0.4756 0.203 0.3677 0.3678 41.4318

Framework versions

  • Transformers 4.36.0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.5
  • Tokenizers 0.15.0