Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Divehi
whisper
Generated from Trainer
Eval Results (legacy)
# Load model directly
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
processor = AutoProcessor.from_pretrained("bunduli/whisper-small-dv-second")
model = AutoModelForSpeechSeq2Seq.from_pretrained("bunduli/whisper-small-dv-second")Quick Links
Whisper Small Dv - Sanchit Gandhi
This model is a fine-tuned version of openai/whisper-small on the Common Voice 13 dataset. It achieves the following results on the evaluation set:
- Loss: 0.1743
- Wer Ortho: 0.6296
- Wer: 0.1351
- Cer: 0.0968
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: 16
- 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: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer | Cer |
|---|---|---|---|---|---|---|
| 0.2081 | 0.8143 | 250 | 0.2399 | 0.7501 | 0.1767 | 0.1249 |
| 0.1206 | 1.6287 | 500 | 0.1743 | 0.6296 | 0.1351 | 0.0968 |
Framework versions
- Transformers 4.46.1
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.20.2
- Downloads last month
- 2
Model tree for bunduli/whisper-small-dv-second
Base model
openai/whisper-smallDataset used to train bunduli/whisper-small-dv-second
Evaluation results
- Wer on Common Voice 13self-reported0.135
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="bunduli/whisper-small-dv-second")