vit-base-lcdoctypev1_session2

This model is a fine-tuned version of nsugianto/vit-base-lcdoctypev1 on the doctype_v1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1231
• Accuracy: 0.9752

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.08	5	0.3652	0.9008
0.2494	0.17	10	0.4354	0.8926
0.2494	0.25	15	0.6027	0.8264
0.4714	0.33	20	0.6016	0.8182
0.4714	0.42	25	0.6495	0.8017
0.6257	0.5	30	0.6561	0.7769
0.6257	0.58	35	0.4359	0.8264
0.3534	0.67	40	0.4082	0.8760
0.3534	0.75	45	0.3042	0.9174
0.2059	0.83	50	0.3919	0.8843
0.2059	0.92	55	0.9259	0.7521
0.3693	1.0	60	0.5527	0.8430
0.3693	1.08	65	0.3412	0.8595
0.3436	1.17	70	0.3949	0.8595
0.3436	1.25	75	0.4402	0.8760
0.361	1.33	80	0.8712	0.7355
0.361	1.42	85	0.4137	0.8430
0.4531	1.5	90	0.2484	0.9256
0.4531	1.58	95	0.3072	0.9008
0.223	1.67	100	0.3936	0.9008
0.223	1.75	105	0.4798	0.8595
0.3396	1.83	110	0.3731	0.8760
0.3396	1.92	115	0.3138	0.9174
0.1516	2.0	120	0.3140	0.9008
0.1516	2.08	125	0.3938	0.8926
0.1709	2.17	130	0.3311	0.9174
0.1709	2.25	135	0.3652	0.9091
0.1193	2.33	140	0.4033	0.9008
0.1193	2.42	145	0.3220	0.8926
0.2134	2.5	150	0.4836	0.8595
0.2134	2.58	155	0.2969	0.9091
0.1314	2.67	160	0.3032	0.9174
0.1314	2.75	165	0.3466	0.8843
0.2197	2.83	170	0.2749	0.8843
0.2197	2.92	175	0.2071	0.9256
0.1748	3.0	180	0.2477	0.9256
0.1748	3.08	185	0.2279	0.9174
0.0949	3.17	190	0.2859	0.9091
0.0949	3.25	195	0.2731	0.9008
0.1901	3.33	200	0.2329	0.9339
0.1901	3.42	205	0.4313	0.9008
0.1897	3.5	210	0.3919	0.9008
0.1897	3.58	215	0.2083	0.9504
0.1213	3.67	220	0.2194	0.9421
0.1213	3.75	225	0.2504	0.9421
0.1225	3.83	230	0.2727	0.9339
0.1225	3.92	235	0.5615	0.8595
0.148	4.0	240	0.3485	0.9091
0.148	4.08	245	0.2324	0.9008
0.1498	4.17	250	0.2689	0.9256
0.1498	4.25	255	0.1947	0.9504
0.0504	4.33	260	0.2275	0.9339
0.0504	4.42	265	0.2790	0.9339
0.0465	4.5	270	0.2727	0.9256
0.0465	4.58	275	0.2949	0.9256
0.0744	4.67	280	0.2705	0.9421
0.0744	4.75	285	0.2787	0.9339
0.1504	4.83	290	0.2361	0.9421
0.1504	4.92	295	0.2886	0.9256
0.1192	5.0	300	0.2602	0.9256
0.1192	5.08	305	0.2183	0.9256
0.0749	5.17	310	0.2774	0.9174
0.0749	5.25	315	0.2491	0.9091
0.0719	5.33	320	0.1902	0.9504
0.0719	5.42	325	0.1915	0.9504
0.061	5.5	330	0.1713	0.9587
0.061	5.58	335	0.2117	0.9504
0.0381	5.67	340	0.2145	0.9587
0.0381	5.75	345	0.2108	0.9504
0.0494	5.83	350	0.2006	0.9504
0.0494	5.92	355	0.1935	0.9504
0.0341	6.0	360	0.1891	0.9669
0.0341	6.08	365	0.1675	0.9669
0.0186	6.17	370	0.1582	0.9669
0.0186	6.25	375	0.1470	0.9669
0.0444	6.33	380	0.1401	0.9752
0.0444	6.42	385	0.1611	0.9669
0.031	6.5	390	0.1923	0.9669
0.031	6.58	395	0.2467	0.9339
0.0415	6.67	400	0.1997	0.9587
0.0415	6.75	405	0.1741	0.9669
0.0396	6.83	410	0.1882	0.9669
0.0396	6.92	415	0.2266	0.9587
0.1419	7.0	420	0.2240	0.9587
0.1419	7.08	425	0.1832	0.9669
0.0401	7.17	430	0.1767	0.9587
0.0401	7.25	435	0.1653	0.9587
0.0435	7.33	440	0.1873	0.9421
0.0435	7.42	445	0.2051	0.9504
0.0368	7.5	450	0.2075	0.9504
0.0368	7.58	455	0.1932	0.9504
0.0144	7.67	460	0.1878	0.9504
0.0144	7.75	465	0.1855	0.9504
0.0245	7.83	470	0.1826	0.9587
0.0245	7.92	475	0.1652	0.9669
0.049	8.0	480	0.1540	0.9669
0.049	8.08	485	0.1566	0.9587
0.0205	8.17	490	0.1390	0.9587
0.0205	8.25	495	0.1231	0.9752
0.0582	8.33	500	0.1489	0.9669
0.0582	8.42	505	0.1636	0.9669
0.0233	8.5	510	0.1738	0.9587
0.0233	8.58	515	0.1830	0.9587
0.026	8.67	520	0.1868	0.9587
0.026	8.75	525	0.1848	0.9587
0.0193	8.83	530	0.1805	0.9669
0.0193	8.92	535	0.1774	0.9669
0.0174	9.0	540	0.1709	0.9669
0.0174	9.08	545	0.1626	0.9587
0.036	9.17	550	0.1482	0.9669
0.036	9.25	555	0.1522	0.9669
0.0248	9.33	560	0.1524	0.9669
0.0248	9.42	565	0.1530	0.9669
0.0097	9.5	570	0.1537	0.9669
0.0097	9.58	575	0.1546	0.9669
0.0154	9.67	580	0.1550	0.9669
0.0154	9.75	585	0.1554	0.9669
0.0208	9.83	590	0.1549	0.9669
0.0208	9.92	595	0.1541	0.9669
0.0167	10.0	600	0.1542	0.9669

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2