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-44
results: []
bart-base-summarization-medical-44
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: 2.1287
- Rouge1: 0.4223
- Rouge2: 0.2251
- Rougel: 0.3572
- Rougelsum: 0.357
- Gen Len: 18.196
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: 44
- 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 |
|---|---|---|---|---|---|---|---|---|
| 2.7031 | 1.0 | 1250 | 2.1999 | 0.413 | 0.2201 | 0.3533 | 0.3528 | 17.704 |
| 2.6144 | 2.0 | 2500 | 2.1644 | 0.4143 | 0.2198 | 0.3521 | 0.3518 | 17.965 |
| 2.5745 | 3.0 | 3750 | 2.1561 | 0.4142 | 0.2169 | 0.3486 | 0.3483 | 18.171 |
| 2.5622 | 4.0 | 5000 | 2.1389 | 0.419 | 0.2222 | 0.3523 | 0.3524 | 18.221 |
| 2.5308 | 5.0 | 6250 | 2.1308 | 0.422 | 0.2255 | 0.3569 | 0.3569 | 18.183 |
| 2.5394 | 6.0 | 7500 | 2.1287 | 0.4223 | 0.2251 | 0.3572 | 0.357 | 18.196 |
Framework versions
- PEFT 0.12.0
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1