metadata
base_model: facebook/bart-base
library_name: peft
license: apache-2.0
metrics:
- rouge
tags:
- generated_from_trainer
model-index:
- name: bart-base-summarization-medical_on_cnn-45
results: []
bart-base-summarization-medical_on_cnn-45
This model is a fine-tuned version of facebook/bart-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 3.3819
- Rouge1: 0.2543
- Rouge2: 0.0961
- Rougel: 0.2009
- Rougelsum: 0.2262
- Gen Len: 18.613
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: 4
- eval_batch_size: 1
- seed: 45
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 2.6997 | 1.0 | 1250 | 3.3755 | 0.2507 | 0.0879 | 0.1956 | 0.2199 | 19.0 |
| 2.5999 | 2.0 | 2500 | 3.3728 | 0.2478 | 0.0901 | 0.1965 | 0.22 | 18.974 |
| 2.5783 | 3.0 | 3750 | 3.3670 | 0.2532 | 0.0929 | 0.199 | 0.2243 | 18.781 |
| 2.5498 | 4.0 | 5000 | 3.3812 | 0.2541 | 0.0947 | 0.2005 | 0.2253 | 18.681 |
| 2.5358 | 5.0 | 6250 | 3.3795 | 0.254 | 0.095 | 0.2006 | 0.2252 | 18.677 |
| 2.5296 | 6.0 | 7500 | 3.3819 | 0.2543 | 0.0961 | 0.2009 | 0.2262 | 18.613 |
Framework versions
- PEFT 0.12.0
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1