Downloads · 30 days
17
3% of all-time downloads
ethiopic-asr-characters
Samuael/ethiopic-asr-characters
ethiopic-asr-characters is a automatic speech recognition model from Samuael. Use it when you need speech turned into text. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
612
Public
Parameters
316M
12.6 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.3 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- Model type
- wav2vec2
- Access
- Public
- Created
- Oct 27, 2024
- Updated
- Dec 7, 2024
- SHA
- 8231a2f5
Datasets
Model card
From the Hugging Face model README
<!-- 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. -->
ethiopic-asr-characters
This model is a fine-tuned version of Samuael/ethiopic-asr-characters on the alffa_amharic dataset. It achieves the following results on the evaluation set:
- Loss: 0.4428
- Wer: 0.2996
- Phoneme Cer: 0.1421
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: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Use 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_steps: 100
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Phoneme Cer |
|---|---|---|---|---|---|
| 0.1559 | 0.2312 | 200 | 0.5540 | 0.3212 | 0.1475 |
| 0.1228 | 0.4624 | 400 | 0.5394 | 0.3142 | 0.1466 |
| 0.069 | 0.6936 | 600 | 0.5525 | 0.3139 | 0.1466 |
| 0.0997 | 0.9249 | 800 | 0.5531 | 0.3118 | 0.1462 |
| 0.0732 | 1.1561 | 1000 | 0.5644 | 0.3196 | 0.1465 |
| 0.1637 | 1.3873 | 1200 | 0.5367 | 0.3185 | 0.1468 |
| 0.1145 | 1.6185 | 1400 | 0.5215 | 0.3180 | 0.1468 |
| 0.1499 | 1.8497 | 1600 | 0.4985 | 0.3141 | 0.1455 |
| 0.1975 | 2.0809 | 1800 | 0.4814 | 0.3114 | 0.1446 |
| 0.2116 | 2.3121 | 2000 | 0.4855 | 0.3085 | 0.1446 |
| 0.2384 | 2.5434 | 2200 | 0.4702 | 0.3083 | 0.1441 |
| 0.4346 | 2.7746 | 2400 | 0.4762 | 0.3063 | 0.1435 |
| 0.3156 | 3.0058 | 2600 | 0.4677 | 0.3044 | 0.1432 |
| 0.5558 | 3.2370 | 2800 | 0.4574 | 0.2993 | 0.1425 |
| 0.2787 | 3.4682 | 3000 | 0.4478 | 0.2989 | 0.1420 |
| 0.3268 | 3.6994 | 3200 | 0.4466 | 0.2978 | 0.1418 |
| 0.3461 | 3.9306 | 3400 | 0.4428 | 0.2996 | 0.1421 |
Framework versions
- Transformers 4.46.2
- Pytorch 2.5.1+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- wav2vec2
- Created
- Oct 27, 2024
- Updated
- Dec 7, 2024