|
--- |
|
license: apache-2.0 |
|
base_model: facebook/bart-large |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- rouge |
|
- wer |
|
model-index: |
|
- name: bart_extractive_1024_1000 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# bart_extractive_1024_1000 |
|
|
|
This model is a fine-tuned version of [facebook/bart-large](https://huggingface.co/facebook/bart-large) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.8802 |
|
- Rouge1: 0.7215 |
|
- Rouge2: 0.4773 |
|
- Rougel: 0.668 |
|
- Rougelsum: 0.668 |
|
- Wer: 0.4137 |
|
- Bleurt: -0.027 |
|
|
|
## 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: 6 |
|
- eval_batch_size: 6 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 2 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Wer | Bleurt | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:------:|:-------:| |
|
| No log | 0.13 | 250 | 1.1362 | 0.6713 | 0.4064 | 0.6113 | 0.6111 | 0.4774 | -0.1118 | |
|
| 2.0454 | 0.27 | 500 | 1.0337 | 0.6869 | 0.4301 | 0.6289 | 0.6288 | 0.4555 | -0.1734 | |
|
| 2.0454 | 0.4 | 750 | 1.0002 | 0.7017 | 0.4465 | 0.6435 | 0.6434 | 0.4467 | -0.357 | |
|
| 1.0987 | 0.53 | 1000 | 0.9747 | 0.7008 | 0.4469 | 0.6423 | 0.6422 | 0.442 | -0.0679 | |
|
| 1.0987 | 0.66 | 1250 | 0.9589 | 0.7092 | 0.456 | 0.6521 | 0.652 | 0.4363 | 0.2669 | |
|
| 1.0418 | 0.8 | 1500 | 0.9551 | 0.704 | 0.4538 | 0.6486 | 0.6485 | 0.4343 | -0.1447 | |
|
| 1.0418 | 0.93 | 1750 | 0.9316 | 0.7096 | 0.4605 | 0.6546 | 0.6544 | 0.4285 | -0.0465 | |
|
| 1.0031 | 1.06 | 2000 | 0.9150 | 0.7129 | 0.4653 | 0.6584 | 0.6583 | 0.4255 | -0.1069 | |
|
| 1.0031 | 1.2 | 2250 | 0.9094 | 0.7119 | 0.4658 | 0.6577 | 0.6576 | 0.4234 | -0.4062 | |
|
| 0.9052 | 1.33 | 2500 | 0.9101 | 0.721 | 0.4736 | 0.6665 | 0.6664 | 0.4206 | 0.2201 | |
|
| 0.9052 | 1.46 | 2750 | 0.8983 | 0.7161 | 0.471 | 0.6619 | 0.6618 | 0.4184 | 0.0117 | |
|
| 0.9045 | 1.6 | 3000 | 0.8917 | 0.7216 | 0.4762 | 0.6675 | 0.6674 | 0.4169 | 0.2346 | |
|
| 0.9045 | 1.73 | 3250 | 0.8906 | 0.7167 | 0.474 | 0.6643 | 0.6642 | 0.4153 | -0.0679 | |
|
| 0.8767 | 1.86 | 3500 | 0.8797 | 0.7232 | 0.4787 | 0.6698 | 0.6697 | 0.4141 | 0.2346 | |
|
| 0.8767 | 1.99 | 3750 | 0.8802 | 0.7215 | 0.4773 | 0.668 | 0.668 | 0.4137 | -0.027 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.38.2 |
|
- Pytorch 2.2.1+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
|
|
