File size: 4,523 Bytes
679f6b8
0af0f86
 
 
 
 
 
 
 
 
 
 
679f6b8
 
0af0f86
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
---
language:
- nl
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper Large V2
  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. -->

# Whisper Large V2

This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2647
- Wer: 9.9296

## 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: 12
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 20
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.6181        | 0.09  | 30   | 0.3443          | 17.6297 |
| 0.3296        | 0.19  | 60   | 0.2921          | 13.0722 |
| 0.3165        | 0.28  | 90   | 0.2711          | 11.8459 |
| 0.2775        | 0.38  | 120  | 0.2677          | 11.4110 |
| 0.2696        | 0.47  | 150  | 0.2570          | 12.1474 |
| 0.2558        | 0.57  | 180  | 0.2544          | 13.9623 |
| 0.272         | 0.66  | 210  | 0.2448          | 19.9809 |
| 0.2696        | 0.76  | 240  | 0.2415          | 12.5359 |
| 0.2668        | 0.85  | 270  | 0.2392          | 11.7154 |
| 0.2558        | 0.95  | 300  | 0.2318          | 12.3097 |
| 0.2108        | 1.04  | 330  | 0.2418          | 15.0639 |
| 0.1339        | 1.14  | 360  | 0.2409          | 13.8318 |
| 0.1326        | 1.23  | 390  | 0.2394          | 16.9165 |
| 0.1357        | 1.33  | 420  | 0.2362          | 11.4371 |
| 0.1278        | 1.42  | 450  | 0.2377          | 12.3503 |
| 0.1485        | 1.52  | 480  | 0.2291          | 12.3097 |
| 0.1263        | 1.61  | 510  | 0.2381          | 12.2547 |
| 0.1269        | 1.71  | 540  | 0.2328          | 10.3528 |
| 0.1437        | 1.8   | 570  | 0.2284          | 11.7415 |
| 0.1314        | 1.9   | 600  | 0.2270          | 11.4661 |
| 0.1282        | 1.99  | 630  | 0.2334          | 11.8807 |
| 0.0646        | 2.09  | 660  | 0.2414          | 11.3617 |
| 0.0655        | 2.18  | 690  | 0.2439          | 11.9503 |
| 0.0658        | 2.28  | 720  | 0.2406          | 11.4197 |
| 0.0573        | 2.37  | 750  | 0.2375          | 11.0747 |
| 0.0584        | 2.47  | 780  | 0.2364          | 10.7848 |
| 0.0639        | 2.56  | 810  | 0.2351          | 11.2197 |
| 0.0537        | 2.66  | 840  | 0.2380          | 10.0803 |
| 0.0538        | 2.75  | 870  | 0.2359          | 10.1180 |
| 0.0551        | 2.85  | 900  | 0.2347          | 10.1847 |
| 0.0613        | 2.94  | 930  | 0.2354          | 10.7616 |
| 0.0453        | 3.04  | 960  | 0.2399          | 13.9130 |
| 0.0248        | 3.13  | 990  | 0.2456          | 11.4139 |
| 0.0246        | 3.23  | 1020 | 0.2553          | 11.3936 |
| 0.0229        | 3.32  | 1050 | 0.2477          | 11.3878 |
| 0.0198        | 3.42  | 1080 | 0.2486          | 10.2137 |
| 0.0229        | 3.51  | 1110 | 0.2491          | 10.2514 |
| 0.021         | 3.61  | 1140 | 0.2478          | 10.4311 |
| 0.0212        | 3.7   | 1170 | 0.2482          | 10.4398 |
| 0.0218        | 3.8   | 1200 | 0.2474          | 10.8892 |
| 0.0225        | 3.89  | 1230 | 0.2442          | 10.3731 |
| 0.0209        | 3.99  | 1260 | 0.2439          | 10.5326 |
| 0.0105        | 4.08  | 1290 | 0.2544          | 10.2948 |
| 0.0089        | 4.18  | 1320 | 0.2614          | 10.3238 |
| 0.008         | 4.27  | 1350 | 0.2627          | 10.0223 |
| 0.0079        | 4.37  | 1380 | 0.2652          | 10.2456 |
| 0.0083        | 4.46  | 1410 | 0.2646          | 10.1267 |
| 0.0073        | 4.56  | 1440 | 0.2619          | 10.0136 |
| 0.0075        | 4.65  | 1470 | 0.2633          | 9.7266  |
| 0.0068        | 4.75  | 1500 | 0.2648          | 9.8281  |
| 0.0074        | 4.84  | 1530 | 0.2645          | 9.8194  |
| 0.0079        | 4.94  | 1560 | 0.2647          | 9.9296  |


### Framework versions

- Transformers 4.38.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.0