Merge branch 'main' of https://huggingface.co/gaduhhartawan/indobart-small
Browse files
README.md
CHANGED
|
@@ -1,3 +1,37 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
+
datasets:
|
| 4 |
+
- id_liputan6
|
| 5 |
+
language:
|
| 6 |
+
- id
|
| 7 |
+
metrics:
|
| 8 |
+
- rouge
|
| 9 |
+
pipeline_tag: summarization
|
| 10 |
+
tags:
|
| 11 |
+
- bart
|
| 12 |
---
|
| 13 |
+
|
| 14 |
+
# indobart-small
|
| 15 |
+
This model is a fine-tuned version of [bart-large-cnn](https://huggingface.co/facebook/bart-large-cnn) on [Liputan6](https://paperswithcode.com/dataset/liputan6) dataset.
|
| 16 |
+
|
| 17 |
+
## Training procedure
|
| 18 |
+
### Training hyperparameters
|
| 19 |
+
- learning_rate: 0.0001
|
| 20 |
+
- train_batch_size: 4
|
| 21 |
+
- eval_batch_size: 4
|
| 22 |
+
- seed: 42
|
| 23 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 24 |
+
- lr_scheduler_type: linear
|
| 25 |
+
- num_epochs: 1
|
| 26 |
+
|
| 27 |
+
### Training results
|
| 28 |
+
|
| 29 |
+
| Training Loss | Epoch | R1 Precision | R1 Recall | R1 Fmeasure | R2 Precision | R2 Recall | R2 Fmeasure | Rl Precision | Rl Recall | Rl Fmeasure |
|
| 30 |
+
|:-------------:|:-----:|:------------:|:---------:|:-----------:|:------------:|:---------:|:-----------:|:------------:|:---------:|:-----------:|
|
| 31 |
+
| 0.3064 | 1.0 | 0.3487 | 0.6043 | 0.4375 | 0.1318 | 0.2613 | 0.1723 | 0.3349 | 0.5833 | 0.4208 |
|
| 32 |
+
|
| 33 |
+
## Framework versions
|
| 34 |
+
- Transformers 4.40.0
|
| 35 |
+
- Pytorch 2.2.1+cu121
|
| 36 |
+
- Datasets 2.19.0
|
| 37 |
+
- Tokenizers 0.19.1
|