learning_rate: 0.0005
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.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: 0.25
num_epochs: 60.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
4342.3143	1.0	972	316.8274	0.5589	0.1405
739.3038	2.0	1944	171.7839	0.2812	0.0688
543.3055	3.0	2916	245.0542	0.2999	0.0837
506.2439	4.0	3888	170.9789	0.2747	0.0711
489.5725	5.0	4860	153.4695	0.2966	0.0707
511.2345	6.0	5832	188.1195	0.2755	0.0731
548.9239	7.0	6804	215.1733	0.3164	0.0861
549.0804	8.0	7776	215.8654	0.3446	0.0941
590.6141	9.0	8748	227.8594	0.3684	0.0957
641.1532	10.0	9720	257.0021	0.3531	0.0998
653.7033	11.0	10692	263.8706	0.3835	0.1073
704.7661	12.0	11664	271.6261	0.4158	0.1212
739.9211	13.0	12636	311.8914	0.4486	0.1328
785.4281	14.0	13608	364.6550	0.5054	0.1502
826.7395	15.0	14580	317.6146	0.4881	0.1474
828.1573	16.0	15552	327.4776	0.4754	0.1461
810.4266	17.0	16524	321.7529	0.4818	0.1440
781.9373	18.0	17496	301.5043	0.4620	0.1341
745.831	19.0	18468	294.4062	0.4483	0.1268
720.9116	20.0	19440	281.9134	0.4185	0.1234
688.3029	21.0	20412	293.3218	0.4098	0.1180
667.9529	22.0	21384	298.2676	0.4274	0.1258
642.8454	23.0	22356	256.3312	0.4221	0.1183
624.7632	24.0	23328	267.7058	0.3898	0.1130
591.3171	25.0	24300	273.8975	0.3670	0.1089
581.4323	26.0	25272	221.8367	0.3508	0.0982
539.7339	27.0	26244	236.9720	0.3364	0.0969
536.5811	28.0	27216	215.0129	0.3269	0.0919
509.0496	29.0	28188	194.6736	0.3308	0.0928
491.9399	30.0	29160	211.2529	0.3198	0.0904
471.1835	31.0	30132	188.4942	0.2939	0.0825
451.6312	32.0	31104	188.2231	0.3014	0.0808
432.8911	33.0	32076	179.4153	0.3059	0.0815
426.9429	34.0	33048	178.9274	0.2877	0.0794
417.0888	35.0	34020	186.1045	0.2901	0.0813
390.4812	36.0	34992	176.1986	0.2750	0.0758
372.8215	37.0	35964	159.8896	0.2745	0.0743
362.2078	38.0	36936	171.5661	0.2874	0.0748
344.0364	39.0	37908	172.1537	0.2561	0.0707
332.623	40.0	38880	168.3665	0.2476	0.0669
312.4819	41.0	39852	152.4227	0.2464	0.0664
299.782	42.0	40824	147.3899	0.2375	0.0654
296.7858	43.0	41796	149.1521	0.2251	0.0624
277.1522	44.0	42768	162.4419	0.2247	0.0616
271.4269	45.0	43740	135.6027	0.2182	0.0595
253.2759	46.0	44712	142.2380	0.2197	0.0599
238.3421	47.0	45684	141.9042	0.2074	0.0558
232.9579	48.0	46656	148.5058	0.2011	0.0543
224.7665	49.0	47628	133.4907	0.1946	0.0537
211.1498	50.0	48600	140.2976	0.1956	0.0527
208.794	51.0	49572	132.7056	0.1867	0.0513
192.3078	52.0	50544	136.9901	0.1821	0.0498
186.0501	53.0	51516	139.0100	0.1829	0.0501
177.6769	54.0	52488	137.1026	0.1756	0.0481
168.7363	55.0	53460	139.9865	0.1753	0.0481
162.5846	56.0	54432	137.0002	0.1699	0.0459
156.3388	57.0	55404	137.2488	0.1687	0.0461
150.741	58.0	56376	135.8838	0.1664	0.0457
146.9406	59.0	57348	137.1219	0.1640	0.0447
149.5945	60.0	58320	136.3459	0.1648	0.0448

Framework versions

Transformers 4.48.3
Pytorch 2.5.1
Datasets 3.2.0
Tokenizers 0.21.0

Priyanship/base_sami_22k_ft_pseudo_widv_ep60_tde_shuffs3

base_sami_22k_ft_pseudo_widv_ep60_tde_shuffs3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Priyanship/
base_sami_22k_ft_pseudo_widv_ep60_tde_shuffs3