File size: 1,402 Bytes
10d35b9
 
 
 
 
 
 
2b7300e
10d35b9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bbd65e0
 
10d35b9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ko
license: apache-2.0
tags:
- hf-asr-leaderboard
- generated_from_trainer
base_model: openai/whisper-small
datasets:
- korean_samll_dataset2
model-index:
- name: korean-small_t3
  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. -->

# korean-small_t3

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the korean_samll_dataset2 dataset.
It achieves the following results on the evaluation set:
- eval_loss: 3.5461
- eval_cer: 235.5384
- eval_runtime: 128.9896
- eval_samples_per_second: 1.558
- eval_steps_per_second: 0.202
- step: 0

## 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: 1e-05
- train_batch_size: 16
- 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: 500
- num_epochs: 3.0
- mixed_precision_training: Native AMP

### Framework versions

- Transformers 4.39.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2