EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-08-10

This model is a fine-tuned version of microsoft/layoutlmv3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 10.1413
• Accuracy: 0.7325
• Exit 0 Accuracy: 0.1725
• Exit 1 Accuracy: 0.2175
• Exit 2 Accuracy: 0.6075
• Exit 3 Accuracy: 0.715
• Exit 4 Accuracy: 0.735

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: 2e-05
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 24
• total_train_batch_size: 48
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Exit 0 Accuracy	Exit 1 Accuracy	Exit 2 Accuracy	Exit 3 Accuracy	Exit 4 Accuracy
No log	0.96	16	16.4817	0.17	0.0825	0.045	0.105	0.0625	0.0625
No log	1.98	33	15.9950	0.2675	0.1	0.1325	0.195	0.1775	0.2425
No log	3.0	50	14.9811	0.475	0.1025	0.1475	0.24	0.29	0.4425
No log	3.96	66	14.0127	0.5675	0.105	0.1425	0.27	0.3975	0.505
No log	4.98	83	13.3047	0.6075	0.125	0.1425	0.3175	0.43	0.595
No log	6.0	100	12.7573	0.6125	0.125	0.1475	0.325	0.495	0.615
No log	6.96	116	12.3656	0.645	0.1175	0.155	0.33	0.5175	0.6375
No log	7.98	133	11.9582	0.6625	0.115	0.16	0.3525	0.5725	0.67
No log	9.0	150	11.6533	0.6825	0.1225	0.16	0.375	0.6	0.7075
No log	9.96	166	11.5143	0.685	0.1525	0.1625	0.38	0.6	0.675
No log	10.98	183	11.3152	0.6625	0.115	0.1625	0.41	0.6225	0.6725
No log	12.0	200	11.0708	0.695	0.11	0.1625	0.425	0.6225	0.7075
No log	12.96	216	11.0412	0.6975	0.1125	0.1575	0.4	0.645	0.685
No log	13.98	233	10.8782	0.7125	0.1425	0.165	0.4275	0.6325	0.7075
No log	15.0	250	10.7282	0.7075	0.115	0.165	0.4225	0.65	0.7175
No log	15.96	266	10.7039	0.695	0.15	0.16	0.4375	0.6375	0.69
No log	16.98	283	10.5455	0.7125	0.13	0.165	0.4375	0.6675	0.715
No log	18.0	300	10.5214	0.7075	0.1275	0.17	0.45	0.6825	0.7075
No log	18.96	316	10.4995	0.715	0.155	0.1725	0.4525	0.68	0.7125
No log	19.98	333	10.3224	0.725	0.1475	0.1825	0.46	0.68	0.7225
No log	21.0	350	10.4247	0.71	0.1425	0.1825	0.4625	0.68	0.71
No log	21.96	366	10.3881	0.705	0.1375	0.1825	0.46	0.66	0.7125
No log	22.98	383	10.3065	0.715	0.1375	0.1875	0.465	0.6925	0.7225
No log	24.0	400	10.1955	0.72	0.145	0.1875	0.4725	0.695	0.7225
No log	24.96	416	10.1607	0.72	0.165	0.19	0.4925	0.7075	0.7175
No log	25.98	433	10.2416	0.72	0.14	0.195	0.48	0.7025	0.7275
No log	27.0	450	10.1321	0.715	0.145	0.1875	0.4925	0.7125	0.72
No log	27.96	466	10.1982	0.7275	0.145	0.1875	0.4875	0.7075	0.73
No log	28.98	483	10.2237	0.72	0.1575	0.19	0.515	0.7	0.7225
10.0174	30.0	500	10.1426	0.7175	0.1675	0.1975	0.5275	0.7125	0.7225
10.0174	30.96	516	10.1056	0.7325	0.14	0.1975	0.515	0.715	0.7325
10.0174	31.98	533	10.1616	0.7225	0.1525	0.195	0.5275	0.7175	0.72
10.0174	33.0	550	10.1053	0.7325	0.1425	0.195	0.525	0.7125	0.7275
10.0174	33.96	566	10.1581	0.7175	0.165	0.2	0.5375	0.71	0.71
10.0174	34.98	583	10.0835	0.7225	0.15	0.2025	0.5375	0.715	0.7225
10.0174	36.0	600	10.1349	0.725	0.1425	0.2	0.5375	0.7025	0.725
10.0174	36.96	616	10.0424	0.7325	0.1625	0.1975	0.545	0.7225	0.735
10.0174	37.98	633	10.0692	0.73	0.155	0.195	0.5525	0.7225	0.74
10.0174	39.0	650	10.0838	0.7325	0.1625	0.1975	0.56	0.7225	0.7375
10.0174	39.96	666	10.1160	0.7275	0.1675	0.1975	0.5575	0.7225	0.725
10.0174	40.98	683	10.0971	0.735	0.1675	0.1975	0.5625	0.7175	0.73
10.0174	42.0	700	10.1207	0.73	0.165	0.2	0.5775	0.715	0.7275
10.0174	42.96	716	10.1448	0.7325	0.175	0.205	0.5775	0.7175	0.73
10.0174	43.98	733	10.0945	0.735	0.1675	0.21	0.5775	0.7175	0.735
10.0174	45.0	750	10.1789	0.73	0.17	0.2175	0.5775	0.7125	0.7275
10.0174	45.96	766	10.1274	0.735	0.175	0.215	0.5875	0.7075	0.735
10.0174	46.98	783	10.1656	0.735	0.155	0.2125	0.5875	0.7125	0.7375
10.0174	48.0	800	10.1557	0.7275	0.16	0.215	0.6025	0.715	0.7325
10.0174	48.96	816	10.1436	0.74	0.165	0.215	0.6025	0.7175	0.735
10.0174	49.98	833	10.1474	0.7325	0.1625	0.215	0.6	0.715	0.735
10.0174	51.0	850	10.1647	0.7275	0.1725	0.2175	0.605	0.7175	0.7325
10.0174	51.96	866	10.1375	0.73	0.1775	0.215	0.6025	0.7125	0.7375
10.0174	52.98	883	10.1458	0.7325	0.1675	0.2175	0.605	0.7125	0.7375
10.0174	54.0	900	10.1527	0.7275	0.175	0.22	0.6025	0.715	0.73
10.0174	54.96	916	10.1349	0.7325	0.175	0.2175	0.6025	0.72	0.735
10.0174	55.98	933	10.1376	0.7325	0.175	0.22	0.6025	0.72	0.7325
10.0174	57.0	950	10.1413	0.7325	0.1725	0.2175	0.6075	0.715	0.7325
10.0174	57.6	960	10.1413	0.7325	0.1725	0.2175	0.6075	0.715	0.735

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3