whisper-ocw – AI Model by sonqwe | AlphaNeural AI

You can deploy this model and start earning money today!

sonqwe/
whisper-ocw

like

0

generated_from_trainer

openai/whisper-base.en

Views

No views yet

whisper-ocw

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

Loss: 0.3009
Wer: 12.6369

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: 0.0001
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
training_steps: 7758
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.9571	0.5	387	0.9503	13.1921
0.5493	1.0	774	0.5891	12.7568
0.3654	1.5	1161	0.4173	12.6168
0.2705	1.99	1548	0.3528	12.5351
0.2589	2.49	1935	0.3262	12.1694
0.2387	2.99	2322	0.3164	12.4052
0.22	3.49	2709	0.3110	12.1956
0.2113	3.99	3096	0.3073	12.1925
0.1963	4.49	3483	0.3052	12.2651
0.1983	4.99	3870	0.3044	12.2852
0.2045	5.49	4257	0.3034	12.2389
0.1865	5.98	4644	0.3025	12.5634
0.1844	6.48	5031	0.3017	12.5835
0.1913	6.98	5418	0.3015	12.5865
0.193	7.48	5805	0.3012	12.5946
0.1951	7.98	6192	0.3009	12.6097
0.1858	8.48	6579	0.3008	12.6198
0.1991	8.98	6966	0.3011	12.6127
0.1767	9.48	7353	0.3009	12.6248
0.1885	9.97	7740	0.3009	12.6369

Framework versions

Transformers 4.37.0.dev0
Pytorch 2.6.0+cu124
Datasets 3.2.0
Tokenizers 0.15.2