You can deploy this model and start earning money today!

nik1509/
telugu_wav2vec_lr_ablation_1e-4

Name: telugu_wav2vec_lr_ablation_1e-4
Rating: 2.5 (8 reviews)

transformers

safetensors

wav2vec2

automatic-speech-recognition

generated_from_trainer

nik1509/telugu_wav2vecmedium_basemodel

finetune

apache-2.0

endpoints_compatible

Views

No views yet

telugu_wav2vec_lr_ablation_1e-4

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: 655.3532
Wer: 0.3968
Cer: 0.1709

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: 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
lr_scheduler_warmup_ratio: 0.1
num_epochs: 15
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
341.2104	0.1736	200	452.5906	0.4194	0.1719
292.7593	0.3472	400	459.1435	0.4117	0.1694
303.5348	0.5208	600	460.0535	0.4130	0.1703
328.5412	0.6944	800	467.8393	0.4085	0.1689
319.0793	0.8681	1000	467.0869	0.4069	0.1692
330.5378	1.0417	1200	480.4087	0.4028	0.1675
303.2146	1.2153	1400	479.1867	0.4094	0.1689
248.5539	1.3889	1600	481.8775	0.4028	0.1687
311.2026	1.5625	1800	476.4391	0.4044	0.1682
291.0741	1.7361	2000	473.0724	0.4105	0.1692
323.9095	1.9097	2200	475.7026	0.4070	0.1687
259.1843	2.0833	2400	496.7509	0.4077	0.1679
323.0709	2.2569	2600	479.3895	0.4094	0.1690
275.3397	2.4306	2800	488.4514	0.4069	0.1683
303.8953	2.6042	3000	491.3173	0.4059	0.1681
337.1545	2.7778	3200	485.3828	0.4079	0.1683
262.3443	2.9514	3400	494.1664	0.4056	0.1685
251.7434	3.125	3600	482.9182	0.4118	0.1699
270.9845	3.2986	3800	498.5337	0.4094	0.1682
295.2738	3.4722	4000	505.9766	0.4034	0.1665
292.3959	3.6458	4200	487.3198	0.4131	0.1717
324.3001	3.8194	4400	491.8690	0.4108	0.1707
289.8097	3.9931	4600	491.1256	0.4109	0.1693
283.513	4.1667	4800	492.2675	0.4091	0.1694
275.6208	4.3403	5000	495.7697	0.4117	0.1708
302.4433	4.5139	5200	504.2882	0.4087	0.1702
305.52	4.6875	5400	518.1114	0.4030	0.1687
296.6145	4.8611	5600	507.6916	0.3993	0.1676
276.3415	5.0347	5800	525.0587	0.4056	0.1696
287.5717	5.2083	6000	515.1569	0.4061	0.1691
248.4744	5.3819	6200	519.6017	0.4009	0.1694
275.757	5.5556	6400	518.4120	0.4049	0.1697
268.587	5.7292	6600	530.8282	0.3998	0.1677
296.1299	5.9028	6800	508.4438	0.4095	0.1712
235.9683	6.0764	7000	530.7183	0.4020	0.1683
296.7235	6.25	7200	525.0239	0.4064	0.1700
276.2289	6.4236	7400	530.1040	0.4096	0.1711
254.7717	6.5972	7600	539.7398	0.4072	0.1700
331.2648	6.7708	7800	530.2902	0.4029	0.1685
263.9338	6.9444	8000	528.5063	0.4022	0.1691
279.3983	7.1181	8200	538.0709	0.4090	0.1710
277.3282	7.2917	8400	546.0734	0.4048	0.1711
217.8157	7.4653	8600	541.1797	0.4037	0.1699
254.1986	7.6389	8800	540.5525	0.4081	0.1702
284.7999	7.8125	9000	544.5749	0.4077	0.1704
226.0705	7.9861	9200	546.8992	0.3998	0.1683
251.6934	8.1597	9400	552.0984	0.4101	0.1702
217.7454	8.3333	9600	559.5846	0.4062	0.1709
279.59	8.5069	9800	558.2065	0.4026	0.1699
237.0119	8.6806	10000	579.6693	0.3973	0.1702
255.0442	8.8542	10200	554.8525	0.3985	0.1699
244.5447	9.0278	10400	578.3720	0.3957	0.1688
207.4985	9.2014	10600	564.1046	0.3978	0.1697
260.5914	9.375	10800	583.9404	0.4010	0.1706
214.1843	9.5486	11000	570.4279	0.3991	0.1694
272.1538	9.7222	11200	589.7668	0.3985	0.1696
244.6975	9.8958	11400	582.3071	0.3982	0.1690
222.3409	10.0694	11600	592.2450	0.3986	0.1695
225.4762	10.2431	11800	589.2662	0.4000	0.1709
245.8159	10.4167	12000	594.8682	0.3975	0.1701
230.3651	10.5903	12200	565.3562	0.4039	0.1740
249.2454	10.7639	12400	581.9191	0.3985	0.1697
211.3718	10.9375	12600	569.6575	0.4046	0.1725
210.5147	11.1111	12800	592.4904	0.4008	0.1709
228.7967	11.2847	13000	599.6378	0.4013	0.1701
210.5554	11.4583	13200	595.6934	0.4025	0.1715
205.4927	11.6319	13400	589.6932	0.3984	0.1717
242.4075	11.8056	13600	605.1076	0.3949	0.1699
211.818	11.9792	13800	600.9894	0.3973	0.1696
203.6863	12.1528	14000	603.4265	0.3982	0.1717
210.4413	12.3264	14200	615.1299	0.3977	0.1722
212.6799	12.5	14400	615.1391	0.3980	0.1714
204.2507	12.6736	14600	619.2817	0.3983	0.1718
198.6503	12.8472	14800	623.1933	0.3944	0.1712
186.8901	13.0208	15000	628.9558	0.4010	0.1721
224.648	13.1944	15200	631.6039	0.3957	0.1709
219.9113	13.3681	15400	629.9046	0.3943	0.1714
181.9767	13.5417	15600	629.7682	0.3940	0.1707
197.348	13.7153	15800	613.8892	0.3945	0.1710
195.0113	13.8889	16000	613.6852	0.4001	0.1730
187.8998	14.0625	16200	648.1019	0.3941	0.1710
173.2886	14.2361	16400	638.4930	0.3987	0.1715
200.1473	14.4097	16600	652.0002	0.3938	0.1709
208.2087	14.5833	16800	632.0565	0.3956	0.1722
206.5025	14.7569	17000	647.6544	0.3956	0.1718
198.8859	14.9306	17200	655.3532	0.3968	0.1709

Framework versions

Transformers 4.52.4
Pytorch 2.8.0.dev20250609+cu118
Datasets 3.6.0
Tokenizers 0.21.1

nik1509/telugu_wav2vec_lr_ablation_1e-4

telugu_wav2vec_lr_ablation_1e-4

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

nik1509/
telugu_wav2vec_lr_ablation_1e-4