speecht5_finetuned_urdu

This model is a fine-tuned version of microsoft/speecht5_tts on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3509

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: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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_steps: 100
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.7169	0.3481	100	0.5856
0.5139	0.6963	200	0.4443
0.4639	1.0418	300	0.4204
0.4627	1.3899	400	0.4093
0.4494	1.7380	500	0.4021
0.4326	2.0836	600	0.3988
0.4388	2.4317	700	0.3928
0.4327	2.7798	800	0.3884
0.4263	3.1253	900	0.3863
0.4252	3.4735	1000	0.3819
0.4237	3.8216	1100	0.3835
0.4153	4.1671	1200	0.3778
0.4176	4.5152	1300	0.3784
0.4145	4.8634	1400	0.3752
0.4142	5.2089	1500	0.3752
0.4097	5.5570	1600	0.3707
0.404	5.9051	1700	0.3712
0.4044	6.2507	1800	0.3690
0.4068	6.5988	1900	0.3705
0.4026	6.9469	2000	0.3676
0.3982	7.2924	2100	0.3684
0.4017	7.6406	2200	0.3678
0.3999	7.9887	2300	0.3655
0.4015	8.3342	2400	0.3657
0.3992	8.6823	2500	0.3642
0.3877	9.0279	2600	0.3642
0.3981	9.3760	2700	0.3640
0.3936	9.7241	2800	0.3637
0.3814	10.0696	2900	0.3633
0.3907	10.4178	3000	0.3612
0.3953	10.7659	3100	0.3608
0.3887	11.1114	3200	0.3606
0.389	11.4595	3300	0.3617
0.389	11.8077	3400	0.3597
0.385	12.1532	3500	0.3589
0.3866	12.5013	3600	0.3580
0.3882	12.8494	3700	0.3593
0.3908	13.1950	3800	0.3577
0.3873	13.5431	3900	0.3567
0.3874	13.8912	4000	0.3565
0.3912	14.2367	4100	0.3596
0.3856	14.5849	4200	0.3582
0.3857	14.9330	4300	0.3567
0.3859	15.2785	4400	0.3564
0.3803	15.6266	4500	0.3551
0.3837	15.9748	4600	0.3580
0.3842	16.3203	4700	0.3557
0.3827	16.6684	4800	0.3545
0.3691	17.0139	4900	0.3569
0.382	17.3621	5000	0.3572
0.377	17.7102	5100	0.3550
0.3683	18.0557	5200	0.3546
0.3779	18.4038	5300	0.3553
0.3805	18.7520	5400	0.3547
0.3799	19.0975	5500	0.3546
0.3836	19.4456	5600	0.3534
0.3763	19.7937	5700	0.3537
0.3774	20.1393	5800	0.3524
0.3754	20.4874	5900	0.3526
0.3777	20.8355	6000	0.3525
0.3791	21.1810	6100	0.3566
0.3729	21.5292	6200	0.3528
0.3771	21.8773	6300	0.3532
0.3757	22.2228	6400	0.3526
0.3783	22.5709	6500	0.3525
0.377	22.9191	6600	0.3522
0.3799	23.2646	6700	0.3553
0.379	23.6127	6800	0.3528
0.3744	23.9608	6900	0.3523
0.3738	24.3064	7000	0.3533
0.3771	24.6545	7100	0.3526
0.3671	25.0	7200	0.3536
0.3745	25.3481	7300	0.3512
0.3765	25.6963	7400	0.3519
0.3691	26.0418	7500	0.3523
0.3717	26.3899	7600	0.3524
0.3759	26.7380	7700	0.3524
0.3635	27.0836	7800	0.3513
0.3775	27.4317	7900	0.3526
0.377	27.7798	8000	0.3525
0.3731	28.1253	8100	0.3513
0.3721	28.4735	8200	0.3514
0.3756	28.8216	8300	0.3518
0.3718	29.1671	8400	0.3512
0.3745	29.5152	8500	0.3523
0.374	29.8634	8600	0.3528
0.3773	30.2089	8700	0.3507
0.3805	30.5570	8800	0.3504
0.3729	30.9051	8900	0.3518
0.3749	31.2507	9000	0.3511
0.3728	31.5988	9100	0.3518
0.3728	31.9469	9200	0.3519
0.3767	32.2924	9300	0.3505
0.3729	32.6406	9400	0.3504
0.3725	32.9887	9500	0.3496
0.37	33.3342	9600	0.3512
0.371	33.6823	9700	0.3510
0.3623	34.0279	9800	0.3516
0.3721	34.3760	9900	0.3509
0.3722	34.7241	10000	0.3509

Framework versions

• Transformers 4.51.1
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1