Pretrained

Finetuned

PyLaia Pretraining Metrics

Training Summary

Total epochs: 52
Best validation CER: 0.001036 (epoch 47)
Best validation WER: 0.004066 (epoch 47)

Full Metrics Table

epoch	tr_cer	tr_loss	tr_wer	va_cer	va_loss	va_wer
0	0.102511	20.5502	0.228909	0.005607	1.44675	0.025887
1	0.013082	2.87236	0.061452	0.003544	0.977254	0.015911
2	0.010093	2.25076	0.048003	0.002787	0.806774	0.012519
3	0.008387	1.89794	0.039916	0.002174	0.694116	0.009469
4	0.007398	1.69734	0.03535	0.002058	0.658268	0.009019
5	0.006802	1.57578	0.032362	0.001917	0.614705	0.008205
6	0.006398	1.48598	0.03044	0.001798	0.596449	0.007565
7	0.005984	1.40728	0.02844	0.003597	1.37793	0.01733
8	0.00573	1.35598	0.027263	0.001649	0.571631	0.006976
9	0.005494	1.30841	0.026087	0.001704	0.609703	0.007205
10	0.005316	1.27343	0.025282	0.001609	0.585576	0.006896
11	0.005129	1.23307	0.024386	0.001624	0.566718	0.006805
12	0.005015	1.201	0.023676	0.001487	0.561116	0.006289
13	0.004933	1.1912	0.02329	0.001504	0.547125	0.006369
14	0.00474	1.15883	0.022434	0.001563	0.55139	0.006674
15	0.004657	1.13046	0.022109	0.00145	0.527566	0.006078
16	0.004509	1.11441	0.021305	0.001435	0.527379	0.006123
17	0.004429	1.0977	0.020912	0.001401	0.509225	0.00589
18	0.004358	1.07066	0.020649	0.001421	0.496447	0.006016
19	0.004249	1.05592	0.019901	0.001297	0.516378	0.005331
20	0.004206	1.04969	0.019817	0.00136	0.487321	0.005667
21	0.004093	1.03316	0.019192	0.001548	0.526055	0.006689
22	0.004057	1.02762	0.019002	0.001376	0.502673	0.005828
23	0.004001	1.01451	0.018822	0.001361	0.512399	0.005709
24	0.00395	1.00391	0.0185	0.001298	0.496007	0.005319
25	0.003864	0.981246	0.018071	0.001279	0.487199	0.005308
26	0.003849	0.984405	0.018042	0.001238	0.489225	0.005063
27	0.003797	0.971805	0.017749	0.00119	0.487824	0.004838
28	0.003787	0.966469	0.017727	0.001231	0.468324	0.005042
29	0.003697	0.955839	0.017186	0.0012	0.45356	0.00479
30	0.00369	0.961032	0.017177	0.00114	0.453792	0.004524
31	0.003773	0.97912	0.017593	0.001265	0.526798	0.004998
32	0.003655	0.951006	0.017039	0.00125	0.53167	0.00518
33	0.003613	0.936915	0.016874	0.001192	0.481596	0.004827
34	0.00356	0.920974	0.01664	0.00119	0.502282	0.004755
35	0.002972	0.772762	0.013666	0.001186	0.466747	0.004863
36	0.002802	0.726356	0.012762	0.001127	0.458169	0.004509
37	0.002763	0.711398	0.012652	0.00109	0.446273	0.004346
38	0.002683	0.693865	0.012302	0.00106	0.457616	0.004171
39	0.002673	0.693615	0.012196	0.001068	0.453765	0.00427
40	0.002621	0.677396	0.011972	0.001098	0.464599	0.004382
41	0.002617	0.672459	0.011962	0.001062	0.462883	0.004217
42	0.00261	0.674766	0.011838	0.001058	0.460468	0.004213
43	0.00258	0.669272	0.011772	0.001047	0.453072	0.004141
44	0.002565	0.660597	0.011639	0.001038	0.452772	0.004078
45	0.002528	0.654716	0.011505	0.001044	0.452606	0.00413
46	0.002501	0.64822	0.01137	0.00106	0.45137	0.004214
47	0.002542	0.653288	0.011591	0.001036	0.449435	0.004066
48	0.00252	0.650962	0.011476	0.001038	0.446807	0.00407
49	0.002543	0.662387	0.011516	0.00105	0.446939	0.004147
50	0.002528	0.654322	0.01149	0.001044	0.445585	0.004111
51	0.002518	0.653832	0.011452	0.001051	0.447358	0.004158

PyLaia Finetuning Metrics

Training Summary

Total epochs: 27
Best validation CER: 0.000869 (epoch 21)
Best validation WER: 0.003693 (epoch 21)

Full Metrics Table

epoch	tr_cer	tr_loss	tr_wer	va_cer	va_loss	va_wer
0	0.183214	35.8187	0.224537	0.001531	0.370508	0.006411
1	0.003867	0.883387	0.018513	0.001255	0.298736	0.005188
2	0.003428	0.769932	0.016496	0.001143	0.245342	0.00476
3	0.003254	0.734276	0.015593	0.001168	0.262834	0.004929
4	0.003137	0.688775	0.01504	0.001151	0.244334	0.004897
5	0.00301	0.667008	0.014394	0.001086	0.230109	0.004562
6	0.002953	0.656692	0.01415	0.001123	0.238079	0.004684
7	0.002937	0.650136	0.014223	0.001034	0.234374	0.004454
8	0.002913	0.644122	0.014053	0.001017	0.233998	0.004383
9	0.002822	0.626512	0.013617	0.001038	0.23499	0.004427
10	0.00291	0.644877	0.014099	0.001179	0.265034	0.005115
11	0.002319	0.512036	0.01107	0.000923	0.203359	0.003937
12	0.002136	0.469168	0.010157	0.0009	0.201167	0.003825
13	0.002081	0.455159	0.00996	0.0009	0.198162	0.003825
14	0.002001	0.433845	0.009498	0.000908	0.197196	0.003886
15	0.001985	0.425972	0.009495	0.000884	0.195152	0.003785
16	0.001951	0.427156	0.009355	0.000874	0.196313	0.003695
17	0.001979	0.431152	0.009395	0.000897	0.194316	0.003843
18	0.00191	0.418722	0.009082	0.000881	0.198212	0.00374
19	0.001922	0.422092	0.009175	0.000906	0.195854	0.003853
20	0.001939	0.410425	0.009221	0.000888	0.19173	0.003807
21	0.001861	0.404768	0.0089	0.000869	0.194746	0.003693
22	0.001848	0.407902	0.008812	0.000908	0.201331	0.00387
23	0.001844	0.40212	0.008813	0.000894	0.193942	0.003816
24	0.001819	0.399957	0.008706	0.000902	0.200013	0.003802
25	0.001825	0.400195	0.00867	0.000909	0.198319	0.003824
26	0.001822	0.395797	0.008696	0.00088	0.193306	0.003716

johnlockejrr/pylaia-yiddish_synth_pytorch_2

Pretrained

Finetuned

PyLaia Pretraining Metrics

Training Summary

Full Metrics Table

PyLaia Finetuning Metrics

Training Summary

Full Metrics Table

Seed this model