Commit
•
0f8a1bc
1
Parent(s):
052be91
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- summarization
|
| 4 |
+
- persian
|
| 5 |
+
- MBart50
|
| 6 |
+
- Abstractive Summarization
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
datasets:
|
| 9 |
+
- xlsum
|
| 10 |
+
model-index:
|
| 11 |
+
- name: mbart-large-50-finetuned-persian
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 16 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 17 |
+
|
| 18 |
+
# mbart-large-50-finetuned-persian
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50) on the xlsum dataset.
|
| 21 |
+
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 4.1932
|
| 23 |
+
- Rouge-1: 26.11
|
| 24 |
+
- Rouge-2: 8.11
|
| 25 |
+
- Rouge-l: 21.09
|
| 26 |
+
- Gen Len: 37.29
|
| 27 |
+
- Bertscore: 71.08
|
| 28 |
+
|
| 29 |
+
## Model description
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Intended uses & limitations
|
| 34 |
+
|
| 35 |
+
More information needed
|
| 36 |
+
|
| 37 |
+
## Training and evaluation data
|
| 38 |
+
|
| 39 |
+
More information needed
|
| 40 |
+
|
| 41 |
+
## Training procedure
|
| 42 |
+
|
| 43 |
+
### Training hyperparameters
|
| 44 |
+
|
| 45 |
+
The following hyperparameters were used during training:
|
| 46 |
+
- learning_rate: 0.0005
|
| 47 |
+
- train_batch_size: 4
|
| 48 |
+
- eval_batch_size: 4
|
| 49 |
+
- seed: 42
|
| 50 |
+
- gradient_accumulation_steps: 8
|
| 51 |
+
- total_train_batch_size: 32
|
| 52 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 53 |
+
- lr_scheduler_type: linear
|
| 54 |
+
- num_epochs: 5
|
| 55 |
+
- label_smoothing_factor: 0.1
|
| 56 |
+
|
| 57 |
+
### Training results
|
| 58 |
+
|
| 59 |
+
| Training Loss | Epoch | Step | Validation Loss | Rouge-1 | Rouge-2 | Rouge-l | Gen Len | Bertscore |
|
| 60 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:-------:|:---------:|
|
| 61 |
+
| 5.5612 | 1.0 | 1476 | 4.5015 | 17.07 | 3.14 | 13.54 | 47.49 | 66.83 |
|
| 62 |
+
| 4.3049 | 2.0 | 2952 | 4.1055 | 22.63 | 5.89 | 18.03 | 40.43 | 69.23 |
|
| 63 |
+
| 3.8154 | 3.0 | 4428 | 3.9822 | 24.57 | 7.15 | 19.74 | 37.35 | 70.36 |
|
| 64 |
+
| 3.3401 | 4.0 | 5904 | 4.0088 | 25.84 | 7.96 | 20.95 | 37.56 | 70.83 |
|
| 65 |
+
| 2.8879 | 5.0 | 7380 | 4.1932 | 26.24 | 8.26 | 21.23 | 37.78 | 71.05 |
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
### Framework versions
|
| 69 |
+
|
| 70 |
+
- Transformers 4.19.1
|
| 71 |
+
- Pytorch 1.11.0+cu113
|
| 72 |
+
- Datasets 2.2.1
|
| 73 |
+
- Tokenizers 0.12.1
|