File size: 1,748 Bytes
3b93375
5c68093
 
 
639cf06
5c68093
 
 
6657e7a
 
 
 
 
 
 
 
 
 
 
3b93375
 
 
6657e7a
3b93375
6657e7a
 
 
3b93375
 
6657e7a
3b93375
6657e7a
 
 
 
 
 
 
 
 
3b93375
 
173623a
3b93375
173623a
 
3b93375
6657e7a
 
 
 
 
 
3b93375
5c68093
3b93375
5c68093
 
 
 
 
 
 
 
 
 
 
 
3b93375
5c68093
3b93375
5c68093
 
 
6657e7a
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
---
license: apache-2.0
tags:
- generated_from_trainer
base_model: openai/whisper-small
model-index:
- name: arazn-whisper-small-v2
  results: []
datasets:
- ahmedheakl/arzen-llm-speech-ds
language:
- en
- ar
metrics:
- bleu
- cer
- wer
library_name: transformers
pipeline_tag: automatic-speech-recognition
---


# Model Card for Model ID

**Please see paper & code for more information:**
- https://github.com/ahmedheakl/arazn-llm
- https://arxiv.org/abs/2406.18120


## Citation

**BibTeX:**
```
@article{heakl2024arzen,
  title={ArzEn-LLM: Code-Switched Egyptian Arabic-English Translation and Speech Recognition Using LLMs},
  author={Heakl, Ahmed and Zaghloul, Youssef and Ali, Mennatullah and Hossam, Rania and Gomaa, Walid},
  journal={arXiv preprint arXiv:2406.18120},
  year={2024}
}
```


## Model Card Authors

- Email: ahmed.heakl@ejust.edu.eg
- Linkedin: https://linkedin.com/in/ahmed-heakl

<!-- 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. -->

# arazn-whisper-small-v2

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the None dataset.

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 1
- mixed_precision_training: Native AMP

### Framework versions

- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2