You can deploy this model and start earning money today!

ibraheemaloran/
whisper_qlora

Name: whisper_qlora
Availability: InStock

MLCommons/peoples_speech

openai/whisper-small

apache-2.0

model-index

Views

No views yet

Whisper Small QLoRa

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

Loss: 0.3030
Wer: 9.6903

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: 0.0002
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.PAGED_ADAMW with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.03
training_steps: 150

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0687	0.7692	10	0.2805	10.8891
0.0714	1.5385	20	0.2771	10.5894
0.0764	2.3077	30	0.2961	10.1898
0.0527	3.0769	40	0.3059	10.0899
0.0406	3.8462	50	0.2839	10.4396
0.058	4.6154	60	0.2898	10.5395
0.0497	5.3846	70	0.2979	9.9900
0.0351	6.1538	80	0.2725	9.8901
0.0595	6.9231	90	0.3203	9.7902
0.0482	7.6923	100	0.3313	10.0400
0.036	8.4615	110	0.3125	9.6903
0.0479	9.2308	120	0.3384	9.6903
0.0424	10.0	130	0.3298	9.6903
0.0497	10.7692	140	0.3171	9.6903
0.0584	11.5385	150	0.3030	9.6903

Framework versions

PEFT 0.18.0
Transformers 4.57.3
Pytorch 2.9.0+cu126
Datasets 4.0.0
Tokenizers 0.22.1

ibraheemaloran/whisper_qlora

Whisper Small QLoRa

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

ibraheemaloran/
whisper_qlora