File size: 2,878 Bytes
86a95c4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ee61125
 
 
 
86a95c4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ee61125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
86a95c4
 
 
 
 
 
 
 
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
---
library_name: transformers
tags:
- generated_from_trainer
metrics:
- accuracy
- bleu
model-index:
- name: gpt2-wikitext
  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. -->

# gpt2-wikitext

This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1673
- Accuracy: 0.4215
- Perplexity: 23.7437
- Bleu: 0.1486

## 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: 0.0001
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Accuracy | Perplexity | Bleu   |
|:-------------:|:------:|:----:|:---------------:|:--------:|:----------:|:------:|
| 6.0809        | 0.2806 | 500  | 5.9591          | 0.1879   | 387.2722   | 0.0323 |
| 5.0659        | 0.5612 | 1000 | 4.9189          | 0.2618   | 136.8551   | 0.0639 |
| 4.338         | 0.8418 | 1500 | 4.2174          | 0.3219   | 67.8594    | 0.0857 |
| 3.9468        | 1.1223 | 2000 | 3.8832          | 0.3532   | 48.5783    | 0.1094 |
| 3.7577        | 1.4029 | 2500 | 3.7060          | 0.3685   | 40.6890    | 0.1226 |
| 3.6483        | 1.6835 | 3000 | 3.5831          | 0.3787   | 35.9858    | 0.1296 |
| 3.5432        | 1.9641 | 3500 | 3.4948          | 0.3875   | 32.9448    | 0.1360 |
| 3.4221        | 2.2447 | 4000 | 3.4280          | 0.3939   | 30.8160    | 0.1306 |
| 3.3602        | 2.5253 | 4500 | 3.3724          | 0.3991   | 29.1478    | 0.1391 |
| 3.3285        | 2.8058 | 5000 | 3.3261          | 0.4038   | 27.8284    | 0.1369 |
| 3.2072        | 3.0864 | 5500 | 3.2882          | 0.4077   | 26.7936    | 0.1447 |
| 3.2002        | 3.3670 | 6000 | 3.2611          | 0.4112   | 26.0792    | 0.1472 |
| 3.1782        | 3.6476 | 6500 | 3.2317          | 0.4138   | 25.3223    | 0.1421 |
| 3.153         | 3.9282 | 7000 | 3.2080          | 0.4164   | 24.7294    | 0.1415 |
| 3.0846        | 4.2088 | 7500 | 3.1915          | 0.4185   | 24.3249    | 0.1470 |
| 3.0469        | 4.4893 | 8000 | 3.1789          | 0.4199   | 24.0205    | 0.1444 |
| 3.0567        | 4.7699 | 8500 | 3.1673          | 0.4215   | 23.7437    | 0.1486 |


### Framework versions

- Transformers 4.49.0
- Pytorch 2.6.0+cu124
- Datasets 3.3.2
- Tokenizers 0.21.0