File size: 2,968 Bytes
b5c5c10
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9959c9d
 
 
 
b5c5c10
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9959c9d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b5c5c10
 
 
 
9959c9d
b5c5c10
 
9959c9d
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
---
license: apache-2.0
base_model: t5-base
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: t5-base-finetuned-xsum
  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. -->

# t5-base-finetuned-xsum

This model is a fine-tuned version of [t5-base](https://huggingface.co/t5-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7758
- Rouge1: 77.9048
- Rouge2: 52.4603
- Rougel: 78.6825
- Rougelsum: 78.3333
- Gen Len: 6.6

## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1  | Rouge2  | Rougel  | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| No log        | 1.0   | 17   | 2.4750          | 49.2456 | 26.8694 | 48.0467 | 48.0189   | 15.2    |
| No log        | 2.0   | 34   | 1.5092          | 68.1774 | 45.2201 | 67.9806 | 68.0505   | 10.2    |
| No log        | 3.0   | 51   | 1.1905          | 73.8611 | 48.5079 | 74.3016 | 74.127    | 7.5     |
| No log        | 4.0   | 68   | 1.0329          | 74.1693 | 46.4048 | 74.7143 | 74.2566   | 7.0     |
| No log        | 5.0   | 85   | 0.9331          | 73.9841 | 45.8016 | 74.5159 | 74.1905   | 6.5333  |
| No log        | 6.0   | 102  | 0.8774          | 74.9841 | 45.8016 | 75.4048 | 75.2222   | 6.5333  |
| No log        | 7.0   | 119  | 0.8377          | 78.2487 | 51.3968 | 79.0212 | 78.6825   | 6.8333  |
| No log        | 8.0   | 136  | 0.8264          | 76.5714 | 50.1349 | 77.3651 | 77.0159   | 6.4667  |
| No log        | 9.0   | 153  | 0.8160          | 76.5714 | 50.1349 | 77.3651 | 77.0159   | 6.4333  |
| No log        | 10.0  | 170  | 0.7945          | 78.709  | 53.4127 | 79.4974 | 79.0132   | 6.6667  |
| No log        | 11.0  | 187  | 0.7846          | 78.709  | 53.4127 | 79.4974 | 79.0132   | 6.6667  |
| No log        | 12.0  | 204  | 0.7794          | 77.9048 | 52.4603 | 78.6825 | 78.3333   | 6.6     |
| No log        | 13.0  | 221  | 0.7783          | 77.9048 | 52.4603 | 78.6825 | 78.3333   | 6.6     |
| No log        | 14.0  | 238  | 0.7764          | 77.9048 | 52.4603 | 78.6825 | 78.3333   | 6.6     |
| No log        | 15.0  | 255  | 0.7758          | 77.9048 | 52.4603 | 78.6825 | 78.3333   | 6.6     |


### Framework versions

- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1