Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Urdu
whisper
Generated from Trainer
Eval Results (legacy)
# Load model directly
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
processor = AutoProcessor.from_pretrained("abdullah090809/whisper-medium-ur")
model = AutoModelForSpeechSeq2Seq.from_pretrained("abdullah090809/whisper-medium-ur")Quick Links
Whisper Medium Ur - Your Name
This model is a fine-tuned version of GogetaBlueMUI/whisper-medium-ur-v3 on the Common Voice 19.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3692
- Wer: 25.0787
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-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1648 | 0.3279 | 250 | 0.3832 | 28.1711 |
| 0.1748 | 0.6557 | 500 | 0.3737 | 30.1650 |
| 0.1887 | 0.9836 | 750 | 0.3587 | 24.8532 |
| 0.132 | 1.3108 | 1000 | 0.3692 | 25.0787 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.5.1+cu121
- Datasets 3.4.1
- Tokenizers 0.21.0
- Downloads last month
- 2
Model tree for abdullah090809/whisper-medium-ur
Dataset used to train abdullah090809/whisper-medium-ur
Evaluation results
- Wer on Common Voice 19.0test set self-reported25.079
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="abdullah090809/whisper-medium-ur")