whisper-large-optimized – AI Model by galelka | AlphaNeural AI

You can deploy this model and start earning money today!

galelka/
whisper-large-optimized

like

0

automatic-speech-recognition

generated_from_trainer

mozilla-foundation/common_voice_11_0

openai/whisper-small

endpoints_compatible

Views

No views yet

Whisper Small Hi - Sanchit Gandhi

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.5304
Wer: 30.4673

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: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
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
0.3345	1.7778	500	0.4103	43.8302
0.1379	3.5556	1000	0.4276	37.6756
0.0612	5.3333	1500	0.4672	32.8039
0.0234	7.1111	2000	0.4806	32.9948
0.0167	8.8889	2500	0.4829	31.0247
0.0074	10.6667	3000	0.5092	34.4762
0.0023	12.4444	3500	0.5247	29.9786
0.0008	14.2222	4000	0.5304	30.4673

Framework versions

Transformers 4.45.2
Pytorch 2.4.1+cu121
Datasets 3.0.1
Tokenizers 0.20.1