EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-08-12_text_vision_only

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: 1.3116
• Accuracy: 0.7775
• Exit 0 Accuracy: 0.065
• Exit 1 Accuracy: 0.09

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
No log	0.96	16	2.6827	0.1275	0.0625	0.04
No log	1.98	33	2.5158	0.25	0.06	0.095
No log	3.0	50	2.3010	0.33	0.0625	0.095
No log	3.96	66	1.9997	0.435	0.055	0.0925
No log	4.98	83	1.7239	0.6	0.06	0.0925
No log	6.0	100	1.4812	0.6175	0.06	0.09
No log	6.96	116	1.2872	0.6875	0.0625	0.09
No log	7.98	133	1.1118	0.74	0.055	0.09
No log	9.0	150	1.0144	0.7425	0.0625	0.09
No log	9.96	166	0.9663	0.7475	0.0575	0.09
No log	10.98	183	0.9532	0.7475	0.0625	0.09
No log	12.0	200	0.9157	0.7525	0.06	0.09
No log	12.96	216	0.8894	0.77	0.06	0.09
No log	13.98	233	0.9460	0.75	0.0625	0.09
No log	15.0	250	1.0019	0.745	0.0625	0.09
No log	15.96	266	0.9059	0.77	0.0625	0.0875
No log	16.98	283	1.0664	0.7325	0.06	0.0875
No log	18.0	300	1.0637	0.74	0.065	0.0875
No log	18.96	316	1.0398	0.7725	0.09	0.085
No log	19.98	333	1.0745	0.775	0.06	0.0875
No log	21.0	350	1.0653	0.78	0.0625	0.0875
No log	21.96	366	1.0705	0.785	0.065	0.0875
No log	22.98	383	1.1014	0.78	0.0725	0.0875
No log	24.0	400	1.1335	0.78	0.0625	0.0875
No log	24.96	416	1.1510	0.775	0.0725	0.0875
No log	25.98	433	1.1528	0.7825	0.0675	0.0875
No log	27.0	450	1.1758	0.7825	0.0625	0.0875
No log	27.96	466	1.1836	0.785	0.07	0.0875
No log	28.98	483	1.1927	0.78	0.0675	0.0875
1.6955	30.0	500	1.2061	0.7825	0.0775	0.0875
1.6955	30.96	516	1.2128	0.7775	0.065	0.0875
1.6955	31.98	533	1.2172	0.7725	0.07	0.0875
1.6955	33.0	550	1.2237	0.775	0.075	0.0875
1.6955	33.96	566	1.2399	0.7775	0.0625	0.0875
1.6955	34.98	583	1.2590	0.78	0.065	0.0875
1.6955	36.0	600	1.2586	0.7825	0.065	0.0875
1.6955	36.96	616	1.2603	0.775	0.0675	0.0875
1.6955	37.98	633	1.2576	0.78	0.065	0.0875
1.6955	39.0	650	1.2698	0.7775	0.075	0.0875
1.6955	39.96	666	1.2775	0.7725	0.075	0.0875
1.6955	40.98	683	1.2769	0.7725	0.07	0.0875
1.6955	42.0	700	1.2769	0.7725	0.0625	0.0875
1.6955	42.96	716	1.2804	0.775	0.0675	0.0875
1.6955	43.98	733	1.2834	0.775	0.065	0.085
1.6955	45.0	750	1.2907	0.7775	0.0675	0.0875
1.6955	45.96	766	1.2968	0.7775	0.0675	0.0875
1.6955	46.98	783	1.2981	0.7775	0.065	0.0875
1.6955	48.0	800	1.3017	0.7775	0.065	0.0875
1.6955	48.96	816	1.3050	0.7775	0.0675	0.09
1.6955	49.98	833	1.3050	0.775	0.07	0.09
1.6955	51.0	850	1.3044	0.775	0.07	0.09
1.6955	51.96	866	1.3057	0.775	0.0675	0.09
1.6955	52.98	883	1.3072	0.7775	0.0675	0.09
1.6955	54.0	900	1.3101	0.7775	0.0675	0.09
1.6955	54.96	916	1.3119	0.7775	0.065	0.09
1.6955	55.98	933	1.3116	0.7775	0.065	0.09
1.6955	57.0	950	1.3115	0.7775	0.065	0.09
1.6955	57.6	960	1.3116	0.7775	0.065	0.09

Framework versions

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