Whisper Small Tamil FLEURS
This model is a fine-tuned version of openai/whisper-small on the google/fleurs ta_in dataset. It achieves the following results on the evaluation set:
- Loss: 0.5390
- Wer: 20.9327
Model description
This model is fine-tuned for 1000 steps on Tamil Fluers data.
- Zero-shot - 35.2 (google/fluers test)
- fine-tune on FLUERS - 20.93 (google/fluers test) (-40%)
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: 32
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0004 | 83.33 | 1000 | 0.5390 | 20.9327 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1.dev0
- Tokenizers 0.13.2
- Downloads last month
- 7
Inference Providers
NEW
This model is not currently available via any of the supported third-party Inference Providers, and
the model is not deployed on the HF Inference API.
Dataset used to train sgangireddy/whisper-small-ta
Space using sgangireddy/whisper-small-ta 1
Evaluation results
- Wer on google/fleurs ta_intest set self-reported20.933