File size: 1,853 Bytes
358d670
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: cc-by-sa-4.0
base_model: p1atdev/t5-base-xlsum-ja
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: Megagon_step3
  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. -->

# Megagon_step3

This model is a fine-tuned version of [p1atdev/t5-base-xlsum-ja](https://huggingface.co/p1atdev/t5-base-xlsum-ja) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1857
- Rouge1: 0.2252
- Rouge2: 0.0901
- Rougel: 0.2243
- Rougelsum: 0.2239
- Gen Len: 10.8153

## 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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
| No log        | 1.0   | 79   | 2.2342          | 0.2695 | 0.1327 | 0.2702 | 0.2686    | 11.036  |
| No log        | 2.0   | 158  | 1.3641          | 0.267  | 0.1222 | 0.2674 | 0.2634    | 10.9775 |
| No log        | 3.0   | 237  | 1.2064          | 0.2307 | 0.099  | 0.2297 | 0.229     | 10.9324 |
| No log        | 4.0   | 316  | 1.1857          | 0.2252 | 0.0901 | 0.2243 | 0.2239    | 10.8153 |


### Framework versions

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