Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Kazakh
whisper
hf-asr-leaderboard
Generated from Trainer
Eval Results (legacy)
# Load model directly
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
processor = AutoProcessor.from_pretrained("Drahokma/whisper-large-v3-kz")
model = AutoModelForSpeechSeq2Seq.from_pretrained("Drahokma/whisper-large-v3-kz", device_map="auto")Quick Links
Whisper Large v3 Kazakh
This model is a fine-tuned version of openai/whisper-large-v3 on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5842
- Wer: 188.0606
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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0003 | 28.5714 | 1000 | 0.4718 | 546.6835 |
| 0.0 | 57.1429 | 2000 | 0.5506 | 175.4264 |
| 0.0 | 85.7143 | 3000 | 0.5751 | 185.3759 |
| 0.0 | 114.2857 | 4000 | 0.5842 | 188.0606 |
Framework versions
- Transformers 4.42.0.dev0
- Pytorch 1.12.0
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 69
Model tree for Drahokma/whisper-large-v3-kz
Base model
openai/whisper-large-v3Dataset used to train Drahokma/whisper-large-v3-kz
Space using Drahokma/whisper-large-v3-kz 1
Evaluation results
- Wer on Common Voice 17.0test set self-reported188.061
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="Drahokma/whisper-large-v3-kz")