telugu_wav2vec_optimablation_adafactor

This model is a fine-tuned version of nik1509/telugu_wav2vecmedium_basemodel on the None dataset. It achieves the following results on the evaluation set:

• Loss: 473.9822
• Wer: 0.4012
• Cer: 0.1666

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
286.3863	1.0	1152	454.0972	0.4208	0.1718
281.9579	2.0	2304	457.6771	0.4132	0.1697
301.2181	3.0	3456	460.8519	0.4090	0.1685
294.6172	4.0	4608	460.5770	0.4081	0.1682
296.4486	5.0	5760	463.2204	0.4066	0.1679
312.1424	6.0	6912	464.7877	0.4059	0.1675
291.5931	7.0	8064	466.9635	0.4043	0.1675
290.4601	8.0	9216	467.6651	0.4039	0.1673
290.0559	9.0	10368	467.9887	0.4021	0.1673
282.5998	10.0	11520	470.1130	0.4024	0.1672
264.7084	11.0	12672	472.3585	0.4019	0.1667
323.1867	12.0	13824	470.8748	0.4025	0.1671
259.911	13.0	14976	472.4793	0.4025	0.1670
286.4286	14.0	16128	474.4262	0.4011	0.1667
295.5091	15.0	17280	473.9822	0.4012	0.1666

Framework versions

• Transformers 4.53.0
• Pytorch 2.8.0.dev20250609+cu118
• Datasets 3.6.0
• Tokenizers 0.21.2