Modelpublic

HouraMor/wh-ft-lr1e6-dtstf5-adm-ga1ba16-st15k-v2-evalstp10-pat20-trainvalch

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes7downloads

Model Card

wh-ft-lr1e6-dtstf5-adm-ga1ba16-st15k-v2-evalstp10-pat20-trainvalch

This model is a fine-tuned version of HouraMor/wh-ft-lr5e6-dtstf5-adm-ga1ba16-st15k-v2-evalstp500-pat5 on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.5650
—Wer: 0.2748
—Cer: 0.2085

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-06
—trainbatchsize: 16
—evalbatchsize: 8
—seed: 42
—optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 250
—training_steps: 5000
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.3641	0.0201	10	0.5640	0.3217	0.2365
0.3049	0.0402	20	0.5639	0.3212	0.2361
0.2036	0.0602	30	0.5635	0.3207	0.2359
0.3869	0.0803	40	0.5630	0.2988	0.2261
0.296	0.1004	50	0.5624	0.2980	0.2255
0.2669	0.1205	60	0.5619	0.3005	0.2276
0.2338	0.1406	70	0.5616	0.3006	0.2275
0.2956	0.1606	80	0.5609	0.2897	0.2190
0.2909	0.1807	90	0.5604	0.2895	0.2188
0.3329	0.2008	100	0.5600	0.2895	0.2190
0.2706	0.2209	110	0.5597	0.2860	0.2161
0.2226	0.2410	120	0.5596	0.2833	0.2149
0.2741	0.2610	130	0.5600	0.2828	0.2148
0.2417	0.2811	140	0.5608	0.2826	0.2145
0.347	0.3012	150	0.5615	0.2811	0.2131
0.3609	0.3213	160	0.5609	0.2826	0.2142
0.2593	0.3414	170	0.5606	0.2809	0.2134
0.3067	0.3614	180	0.5605	0.2764	0.2092
0.1497	0.3815	190	0.5607	0.2741	0.2078
0.2763	0.4016	200	0.5621	0.2724	0.2064
0.2473	0.4217	210	0.5630	0.2724	0.2067
0.2224	0.4418	220	0.5634	0.2738	0.2080
0.2454	0.4618	230	0.5643	0.2764	0.2108
0.1783	0.4819	240	0.5658	0.2760	0.2109
0.2915	0.5020	250	0.5669	0.2756	0.2103
0.3192	0.5221	260	0.5662	0.2739	0.2078
0.3532	0.5422	270	0.5650	0.2748	0.2085

Framework versions

—Transformers 4.55.2
—Pytorch 2.7.0+cu118
—Datasets 2.21.0
—Tokenizers 0.21.4