Downloads · 30 days
2
5% of all-time downloads
isizulu-w2v-bert-2.0
kesbeast23/isizulu-w2v-bert-2.0
isizulu-w2v-bert-2.0 is a automatic speech recognition model from kesbeast23. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
42
Public
Parameters
581M
2.3 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors2.3 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- License
- mit
- Model type
- wav2vec2-bert
- Access
- Public
- Created
- Nov 26, 2025
- Updated
- Nov 27, 2025
- SHA
- 29251c56
Base models
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. -->
isizulu-w2v-bert-2.0
This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6863
- Wer: 0.5031
- Cer: 0.1080
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: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 500
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 9.8238 | 16.004 | 200 | 8.9513 | 1.0 | 1.0 |
| 5.4145 | 33.002 | 400 | 4.4867 | 1.0 | 1.0 |
| 3.2793 | 49.006 | 600 | 3.0346 | 1.0 | 1.0 |
| 3.1853 | 66.004 | 800 | 2.9150 | 1.0 | 1.0 |
| 1.3089 | 83.002 | 1000 | 1.1980 | 0.5848 | 0.1345 |
| 0.79 | 99.006 | 1200 | 0.8875 | 0.5482 | 0.1231 |
| 0.5796 | 116.004 | 1400 | 0.7594 | 0.5189 | 0.1135 |
| 0.4415 | 133.002 | 1600 | 0.6984 | 0.5071 | 0.1090 |
| 0.3758 | 149.006 | 1800 | 0.6879 | 0.5048 | 0.1086 |
| 0.3469 | 166.004 | 2000 | 0.6863 | 0.5031 | 0.1080 |
Framework versions
- Transformers 4.57.3
- Pytorch 2.9.1+cu128
- Datasets 4.4.1
- Tokenizers 0.22.1
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- wav2vec2-bert
- License
- mit
- Created
- Nov 26, 2025
- Updated
- Nov 27, 2025