File size: 3,208 Bytes
e674497
 
 
 
 
 
 
 
 
 
 
 
 
 
 
aa2cb5c
 
 
 
 
 
 
 
 
 
 
 
e674497
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
aa2cb5c
 
 
 
 
 
 
 
 
 
 
 
e674497
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
---
license: apache-2.0
base_model: sshleifer/distilbart-xsum-12-6
tags:
- generated_from_trainer
model-index:
- name: trained-distilbart-abs-0807
  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. -->

# trained-distilbart-abs-0807

This model is a fine-tuned version of [sshleifer/distilbart-xsum-12-6](https://huggingface.co/sshleifer/distilbart-xsum-12-6) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.4184
- Rouge/rouge1: 0.0185
- Rouge/rouge2: 0.0088
- Rouge/rougel: 0.0152
- Rouge/rougelsum: 0.016
- Bertscore/bertscore-precision: 0.0404
- Bertscore/bertscore-recall: 0.04
- Bertscore/bertscore-f1: 0.0402
- Meteor: 0.0163
- Gen Len: 80.0

## 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: 4
- seed: 42
- 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 | Rouge/rouge1 | Rouge/rouge2 | Rouge/rougel | Rouge/rougelsum | Bertscore/bertscore-precision | Bertscore/bertscore-recall | Bertscore/bertscore-f1 | Meteor | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------------:|:------------:|:------------:|:---------------:|:-----------------------------:|:--------------------------:|:----------------------:|:------:|:-------:|
| 2.105         | 1.0   | 220  | 2.0884          | 0.4545       | 0.2049       | 0.3881       | 0.3905          | 0.8969                        | 0.88                       | 0.8882                 | 0.3923 | 80.0    |
| 1.8823        | 2.0   | 440  | 2.0066          | 0.3453       | 0.1575       | 0.2965       | 0.2984          | 0.6632                        | 0.6547                     | 0.6587                 | 0.2949 | 80.0    |
| 1.4089        | 3.0   | 660  | 2.0717          | 0.0768       | 0.0337       | 0.0637       | 0.0639          | 0.1559                        | 0.1535                     | 0.1547                 | 0.0667 | 80.0    |
| 1.0687        | 4.0   | 880  | 2.1627          | 0.0125       | 0.0048       | 0.0104       | 0.0114          | 0.0322                        | 0.0317                     | 0.0319                 | 0.0118 | 80.0    |
| 0.7445        | 5.0   | 1100 | 2.2927          | 0.0402       | 0.0177       | 0.0332       | 0.0332          | 0.0815                        | 0.0809                     | 0.0812                 | 0.0374 | 80.0    |
| 0.7619        | 6.0   | 1320 | 2.4184          | 0.0185       | 0.0088       | 0.0152       | 0.016           | 0.0404                        | 0.04                       | 0.0402                 | 0.0163 | 80.0    |


### Framework versions

- Transformers 4.41.2
- Pytorch 2.1.2
- Datasets 2.19.2
- Tokenizers 0.19.1