Pretrained

PyLaia Training Metrics

Training Summary

Total epochs: 76
Best validation CER: 0.005937 (epoch 65)
Best validation WER: 0.016457 (epoch 62)

Full Metrics Table

epoch	tr_cer	tr_wer	train_loss_epoch	train_loss_step	va_cer	va_wer	val_loss
0	0.129245	0.31886	39.4692	12.3066	0.016895	0.052086	5.70541
1	0.027248	0.109056	8.92865	11.5946	0.012966	0.043093	4.50416
2	0.02189	0.086051	7.20965	2.85532	0.0108	0.029745	3.50374
3	0.01887	0.074292	6.18997	7.53472	0.009848	0.028535	3.24674
4	0.017496	0.067933	5.74712	2.76993	0.01269	0.031484	4.38984
5	0.016812	0.063935	5.5235	4.21415	0.00976	0.0278	3.24001
6	0.015686	0.060124	5.10744	1.70046	0.008806	0.02412	2.912
7	0.01477	0.056549	4.81694	0.78812	0.008375	0.022697	2.72714
8	0.014652	0.055727	4.73216	1.51364	0.009163	0.02446	2.85291
9	0.013993	0.053153	4.51401	8.49484	0.008506	0.022889	2.64098
10	0.013825	0.051805	4.46963	17.547	0.009276	0.024612	2.9348
11	0.013776	0.050979	4.4778	1.46055	0.007802	0.02094	2.4691
12	0.013263	0.04988	4.29078	12.1256	0.008381	0.021463	2.56431
13	0.012958	0.048507	4.16489	0.517483	0.009516	0.025627	2.99656
14	0.012718	0.047834	4.08675	15.1593	0.007643	0.020674	2.38766
15	0.012782	0.047346	4.14667	11.5345	0.008414	0.022718	2.53537
16	0.013121	0.04756	4.28663	0.704873	0.008986	0.022636	2.85818
17	0.012989	0.047326	4.20438	0.226486	0.007613	0.019826	2.36213
18	0.012409	0.045543	3.986	8.26527	0.008039	0.021331	2.44117
19	0.012467	0.045933	4.01241	1.33168	0.007981	0.020712	2.5643
20	0.012641	0.046513	4.09855	1.48213	0.008311	0.021949	2.48846
21	0.012292	0.045097	3.96557	1.26379	0.007863	0.020584	2.38992
22	0.012018	0.044462	3.863	23.7959	0.008187	0.020854	2.48543
23	0.011232	0.040438	3.53586	0.883133	0.007534	0.020268	2.27983
24	0.010547	0.037563	3.28549	0.167232	0.007306	0.020027	2.25233
25	0.010414	0.037	3.27071	0.371634	0.007236	0.020221	2.22727
26	0.010176	0.036015	3.17962	1.32554	0.007038	0.019737	2.18342
27	0.010005	0.035514	3.13593	1.54786	0.006995	0.019352	2.15701
28	0.009769	0.0347	3.03312	3.60537	0.006651	0.018249	2.08475
29	0.009681	0.034254	3.00317	1.65505	0.006729	0.018778	2.072
30	0.009502	0.033844	2.94229	0.635581	0.00678	0.018843	2.10965
31	0.009577	0.033859	2.94446	0.213813	0.006994	0.01945	2.14548
32	0.00963	0.033855	2.95764	0.406756	0.006863	0.01937	2.09923
33	0.009445	0.033263	2.95371	0.665865	0.006856	0.018825	2.08154
34	0.009255	0.032772	2.86192	0.996484	0.006557	0.018038	2.00478
35	0.009224	0.032645	2.8533	9.40311	0.006938	0.019332	2.10517
36	0.009375	0.03285	2.89019	2.9548	0.006606	0.018241	2.02943
37	0.009225	0.032457	2.84777	12.7469	0.007089	0.019089	2.14697
38	0.009138	0.031929	2.8019	0.16169	0.006444	0.017973	1.96117
39	0.009054	0.031969	2.79038	14.2437	0.006723	0.018281	1.99717
40	0.009104	0.031877	2.79368	8.42458	0.006686	0.018588	2.0263
41	0.009042	0.031731	2.79296	0.014788	0.006775	0.019004	2.04582
42	0.008958	0.031392	2.75503	0.895635	0.006447	0.018211	1.9721
43	0.008843	0.031237	2.73262	5.40016	0.006335	0.017754	1.94061
44	0.0089	0.031457	2.73986	0.01812	0.00648	0.017889	1.96678
45	0.008895	0.031266	2.7319	0.235641	0.006433	0.017983	1.95497
46	0.008668	0.030706	2.68418	3.64022	0.006313	0.01754	1.91343
47	0.008705	0.030696	2.68547	9.00642	0.006597	0.018216	1.96827
48	0.008727	0.030832	2.68516	0.309562	0.006333	0.017886	1.94401
49	0.008538	0.030283	2.64099	0.014653	0.006258	0.017683	1.92426
50	0.008553	0.030119	2.62904	0.243859	0.006267	0.01731	1.91222
51	0.00859	0.030244	2.64838	0.788468	0.006256	0.017584	1.9147
52	0.008542	0.030204	2.61293	1.36889	0.006164	0.017457	1.90145
53	0.008511	0.030127	2.60048	3.81542	0.006275	0.017553	1.90049
54	0.008507	0.030023	2.58761	0.250174	0.006136	0.017044	1.87604
55	0.008361	0.029653	2.57442	0.036428	0.006122	0.017115	1.86923
56	0.008348	0.029541	2.57007	0.167346	0.006051	0.017036	1.84211
57	0.00835	0.029681	2.60374	0.025574	0.005967	0.01684	1.83604
58	0.008356	0.029606	2.56192	0.150598	0.006249	0.017518	1.89512
59	0.008439	0.029808	2.57937	0.088004	0.006153	0.017248	1.89009
60	0.008303	0.029524	2.55604	11.6615	0.006058	0.017358	1.86274
61	0.008235	0.029226	2.53185	0.484483	0.006156	0.017041	1.90813
62	0.008354	0.029538	2.56148	9.77606	0.005976	0.016457	1.8516
63	0.008138	0.028867	2.49159	0.092906	0.006003	0.016879	1.84359
64	0.008134	0.028773	2.47939	11.7722	0.005976	0.016716	1.83542
65	0.008106	0.028679	2.48362	8.10345	0.005937	0.016689	1.82221
66	0.008114	0.028817	2.46834	0.496928	0.005938	0.016746	1.82858
67	0.008079	0.02861	2.47971	2.91834	0.00596	0.016791	1.82926
68	0.00809	0.028662	2.47953	4.82707	0.005963	0.016743	1.82742
69	0.008074	0.028644	2.47506	1.08142	0.005979	0.016915	1.82871
70	0.008035	0.028538	2.45257	0.458501	0.005944	0.016721	1.83211
71	0.00808	0.028751	2.46476	0.604646	0.005949	0.016764	1.83054
72	0.008029	0.028533	2.45826	9.35829	0.005937	0.016716	1.82579
73	0.008038	0.028501	2.44883	1.12761	0.005954	0.016793	1.82871
74	0.008088	0.028739	2.46324	0.010121	0.005958	0.016824	1.82953
75	0.008065	0.028638	2.45915	0.812815	0.005949	0.016781	1.82817

johnlockejrr/pylaia-heb_synth_lines_pytorch_2

Pretrained

PyLaia Training Metrics

Training Summary

Full Metrics Table

为此模型做种