File size: 3,168 Bytes
4f35d33
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
74
75
76
77
78
79
80
81
82
---
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