hypa-intelligence/hypa-whisper-asr-finetuned-streaming-real-data-20260324_113403

Whisper Small Native Multilingual(streamed) - EN/TIV/FR/IG/AN/EFI/NUP/IBB/EGO

This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.4826
• Wer: 0.7513

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: 1
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• 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: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.7441	0.125	500	1.0451	0.5637
3.3957	0.25	1000	1.1103	0.6415
2.7228	0.375	1500	1.3163	0.8318
9.5705	0.5	2000	1.4821	0.7867
3.5521	0.625	2500	1.6896	0.8458
3.6400	0.75	3000	1.5492	0.6993
4.6294	0.875	3500	1.5660	0.7215
3.6829	1.0	4000	1.4826	0.7513

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2