File size: 1,952 Bytes
b8ea4e5
cc9d518
 
 
b8ea4e5
 
cc9d518
b8ea4e5
 
 
cc9d518
b8ea4e5
 
 
 
 
cc9d518
 
 
b8ea4e5
 
 
 
 
 
9240b36
 
 
 
 
 
 
 
b8ea4e5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9240b36
b8ea4e5
 
 
 
 
 
 
 
cc9d518
 
 
 
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
---
base_model: openai/whisper-small
datasets:
- mozilla-foundation/common_voice_17_0
language:
- ps
library_name: transformers.js
license: apache-2.0
tags:
- generated_from_trainer
- onnx
model-index:
- name: Whisper Small PS - Hanif Rahman
  results: []
---

https://huggingface.co/ihanif/whisper-test with ONNX weights to be compatible with Transformers.js.  


<!-- 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 Small PS - Hanif Rahman

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 17.0 dataset.
It achieves the following results on the evaluation set:
- eval_loss: 0.7573
- eval_wer: 46.1819
- eval_runtime: 395.7975
- eval_samples_per_second: 1.294
- eval_steps_per_second: 0.162
- epoch: 5.7143
- step: 2600

## 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
- training_steps: 4000
- mixed_precision_training: Native AMP

### Framework versions

- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0
  
---
  
Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).