EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-05-17_gofor50

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: 1.4441
• Accuracy: 0.7
• Exit 0 Accuracy: 0.15
• Exit 1 Accuracy: 0.1775
• Exit 2 Accuracy: 0.515
• Exit 3 Accuracy: 0.705
• Exit 4 Accuracy: 0.7025

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: 32
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

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	12	2.6817	0.1475	0.0825	0.0525	0.1225	0.0625	0.0625
No log	2.0	25	2.5448	0.235	0.0875	0.1125	0.0875	0.12	0.0625
No log	2.96	37	2.3948	0.2825	0.105	0.14	0.205	0.21	0.0625
No log	4.0	50	2.4502	0.3125	0.1125	0.1425	0.2	0.1725	0.22
No log	4.96	62	2.1482	0.385	0.11	0.145	0.2925	0.3275	0.275
No log	6.0	75	1.8840	0.465	0.1175	0.1475	0.33	0.4075	0.4925
No log	6.96	87	1.6813	0.5725	0.1225	0.16	0.3475	0.4475	0.55
No log	8.0	100	1.5199	0.605	0.13	0.1575	0.3375	0.5	0.585
No log	8.96	112	1.4031	0.6125	0.1325	0.1575	0.365	0.5325	0.6025
No log	10.0	125	1.3312	0.62	0.13	0.1575	0.3825	0.565	0.62
No log	10.96	137	1.2133	0.6575	0.1175	0.1575	0.385	0.605	0.6275
No log	12.0	150	1.1805	0.68	0.135	0.1675	0.38	0.6225	0.645
No log	12.96	162	1.1987	0.6625	0.12	0.165	0.4075	0.63	0.645
No log	14.0	175	1.1304	0.6975	0.12	0.17	0.41	0.6375	0.6725
No log	14.96	187	1.1333	0.695	0.125	0.1675	0.42	0.655	0.675
No log	16.0	200	1.1373	0.6975	0.1425	0.1725	0.4375	0.655	0.6825
No log	16.96	212	1.0988	0.7125	0.1475	0.165	0.44	0.66	0.6975
No log	18.0	225	1.1830	0.7	0.125	0.165	0.455	0.675	0.69
No log	18.96	237	1.1827	0.685	0.13	0.1575	0.45	0.665	0.6825
No log	20.0	250	1.2335	0.695	0.125	0.1575	0.455	0.6675	0.685
No log	20.96	262	1.2793	0.6625	0.135	0.1575	0.4475	0.665	0.6775
No log	22.0	275	1.2136	0.71	0.14	0.155	0.4675	0.6775	0.6925
No log	22.96	287	1.1790	0.7025	0.135	0.16	0.4725	0.6725	0.6975
No log	24.0	300	1.2636	0.6825	0.125	0.1625	0.4375	0.6675	0.6825
No log	24.96	312	1.2949	0.6775	0.1225	0.165	0.46	0.6625	0.685
No log	26.0	325	1.2979	0.685	0.135	0.1625	0.455	0.69	0.6925
No log	26.96	337	1.3059	0.6825	0.1375	0.1675	0.475	0.675	0.69
No log	28.0	350	1.3239	0.6975	0.1375	0.17	0.4725	0.68	0.705
No log	28.96	362	1.3181	0.7	0.1325	0.1725	0.465	0.6725	0.695
No log	30.0	375	1.3978	0.68	0.1375	0.1675	0.4825	0.6825	0.68
No log	30.96	387	1.2876	0.715	0.14	0.175	0.49	0.6925	0.7125
No log	32.0	400	1.3233	0.6975	0.1475	0.1725	0.4875	0.6975	0.705
No log	32.96	412	1.3607	0.7	0.135	0.1725	0.4925	0.6925	0.7075
No log	34.0	425	1.3728	0.6925	0.1525	0.1725	0.4925	0.685	0.7025
No log	34.96	437	1.3931	0.695	0.1425	0.1725	0.49	0.695	0.7
No log	36.0	450	1.3642	0.7125	0.145	0.1725	0.505	0.6875	0.71
No log	36.96	462	1.3807	0.705	0.1475	0.17	0.4975	0.6875	0.695
No log	38.0	475	1.4171	0.7025	0.1575	0.17	0.505	0.6975	0.695
No log	38.96	487	1.3868	0.71	0.155	0.17	0.4975	0.7025	0.705
0.6814	40.0	500	1.3982	0.7075	0.1425	0.17	0.51	0.6975	0.705
0.6814	40.96	512	1.3898	0.7075	0.1475	0.17	0.51	0.6975	0.6975
0.6814	42.0	525	1.4277	0.7025	0.145	0.17	0.515	0.705	0.705
0.6814	42.96	537	1.4381	0.7025	0.145	0.1725	0.51	0.7	0.7025
0.6814	44.0	550	1.4374	0.7025	0.15	0.175	0.5125	0.7	0.7
0.6814	44.96	562	1.4260	0.6975	0.1475	0.175	0.5125	0.7025	0.7
0.6814	46.0	575	1.4408	0.7	0.1475	0.1775	0.5125	0.7025	0.7
0.6814	46.96	587	1.4483	0.7	0.15	0.1775	0.515	0.705	0.7
0.6814	48.0	600	1.4441	0.7	0.15	0.1775	0.515	0.705	0.7025

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 1.12.1+cu113
• Datasets 2.12.0
• Tokenizers 0.12.1