EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-10-02_lte_txt_visual_conc_enc_all

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: 1.6250
• Accuracy: 0.7025
• Exit 0 Accuracy: 0.065
• Exit 1 Accuracy: 0.48
• Exit 2 Accuracy: 0.57
• Exit 3 Accuracy: 0.6425
• Exit 4 Accuracy: 0.6875
• Exit 5 Accuracy: 0.71
• Exit 6 Accuracy: 0.715
• Exit 7 Accuracy: 0.705
• Exit 8 Accuracy: 0.71
• Exit 9 Accuracy: 0.7075
• Exit 10 Accuracy: 0.7075
• Exit 11 Accuracy: 0.705
• Exit 12 Accuracy: 0.7075

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	Exit 5 Accuracy	Exit 6 Accuracy	Exit 7 Accuracy	Exit 8 Accuracy	Exit 9 Accuracy	Exit 10 Accuracy	Exit 11 Accuracy	Exit 12 Accuracy
No log	0.96	16	2.6844	0.1425	0.0725	0.0875	0.0625	0.0625	0.0625	0.07	0.115	0.065	0.0625	0.0625	0.0625	0.0625	0.105
No log	1.98	33	2.5092	0.2425	0.0625	0.075	0.065	0.0625	0.065	0.17	0.13	0.1325	0.0625	0.0625	0.0625	0.0625	0.0875
No log	3.0	50	2.3693	0.2975	0.0575	0.125	0.105	0.1125	0.1075	0.1825	0.1825	0.1825	0.085	0.0675	0.0625	0.085	0.12
No log	3.96	66	2.1547	0.39	0.065	0.1725	0.12	0.1325	0.1175	0.2925	0.2525	0.2725	0.095	0.0625	0.065	0.09	0.145
No log	4.98	83	1.9916	0.45	0.0625	0.2375	0.125	0.1425	0.115	0.3075	0.2825	0.31	0.125	0.08	0.0825	0.1225	0.17
No log	6.0	100	1.8193	0.475	0.0625	0.2725	0.1975	0.1625	0.17	0.405	0.355	0.3775	0.155	0.105	0.1375	0.1225	0.25
No log	6.96	116	1.6391	0.5675	0.06	0.2825	0.2575	0.2175	0.2325	0.435	0.42	0.4125	0.2275	0.1175	0.165	0.125	0.2225
No log	7.98	133	1.5333	0.5975	0.0575	0.2575	0.3	0.3025	0.34	0.445	0.435	0.495	0.3425	0.26	0.1825	0.1325	0.25
No log	9.0	150	1.4413	0.5975	0.06	0.245	0.3075	0.345	0.395	0.4675	0.455	0.5075	0.4375	0.3325	0.2025	0.1375	0.305
No log	9.96	166	1.3251	0.62	0.06	0.3025	0.335	0.3575	0.4525	0.51	0.5275	0.55	0.5125	0.4725	0.2575	0.1875	0.375
No log	10.98	183	1.2767	0.6275	0.07	0.3375	0.345	0.41	0.48	0.5025	0.5125	0.565	0.565	0.555	0.3875	0.2075	0.3625
No log	12.0	200	1.1751	0.675	0.07	0.3425	0.3425	0.44	0.48	0.5275	0.5525	0.5925	0.5575	0.575	0.47	0.26	0.4475
No log	12.96	216	1.2128	0.65	0.0675	0.3375	0.3675	0.4225	0.5125	0.505	0.5325	0.5925	0.5825	0.6125	0.525	0.34	0.48
No log	13.98	233	1.1501	0.6825	0.07	0.35	0.4	0.485	0.5475	0.5425	0.5475	0.5975	0.6125	0.6325	0.5925	0.4425	0.56
No log	15.0	250	1.1550	0.67	0.0675	0.3625	0.425	0.505	0.545	0.5575	0.5875	0.625	0.6	0.6375	0.5975	0.4875	0.5675
No log	15.96	266	1.1341	0.6825	0.0675	0.375	0.445	0.515	0.5575	0.5775	0.61	0.65	0.605	0.635	0.6225	0.54	0.6075
No log	16.98	283	1.1208	0.7025	0.0625	0.385	0.4325	0.5275	0.5825	0.61	0.6275	0.6475	0.635	0.6275	0.62	0.5525	0.6425
No log	18.0	300	1.1508	0.6975	0.065	0.3875	0.4225	0.525	0.5875	0.595	0.6225	0.65	0.6425	0.6225	0.64	0.58	0.665
No log	18.96	316	1.1723	0.7	0.0625	0.3925	0.4475	0.5525	0.585	0.615	0.6475	0.6625	0.655	0.6425	0.65	0.5975	0.65
No log	19.98	333	1.1968	0.6875	0.0675	0.3875	0.465	0.555	0.61	0.6225	0.6325	0.65	0.65	0.645	0.66	0.6325	0.6725
No log	21.0	350	1.2006	0.7	0.0675	0.405	0.4725	0.5625	0.62	0.6225	0.655	0.67	0.66	0.6675	0.69	0.6575	0.6925
No log	21.96	366	1.2073	0.6925	0.0675	0.4175	0.485	0.56	0.62	0.625	0.6425	0.6775	0.6625	0.6825	0.6975	0.705	0.695
No log	22.98	383	1.2901	0.6875	0.07	0.4125	0.4875	0.56	0.6325	0.6425	0.665	0.6725	0.6825	0.69	0.69	0.705	0.695
No log	24.0	400	1.3072	0.69	0.0625	0.42	0.475	0.5675	0.635	0.645	0.665	0.68	0.68	0.6725	0.68	0.69	0.69
No log	24.96	416	1.3590	0.6875	0.0625	0.4375	0.4875	0.5675	0.635	0.65	0.6725	0.6825	0.6825	0.6925	0.69	0.6975	0.695
No log	25.98	433	1.4236	0.6825	0.065	0.435	0.4825	0.565	0.63	0.6325	0.67	0.67	0.67	0.6675	0.6775	0.6875	0.685
No log	27.0	450	1.3880	0.6925	0.065	0.4325	0.5	0.59	0.645	0.65	0.6775	0.6775	0.6825	0.6875	0.68	0.685	0.695
No log	27.96	466	1.3898	0.705	0.0625	0.4325	0.5075	0.6	0.655	0.685	0.68	0.6975	0.7	0.6925	0.685	0.705	0.7125
No log	28.98	483	1.3835	0.7075	0.0625	0.4325	0.5025	0.5925	0.6675	0.67	0.69	0.7025	0.7	0.705	0.6975	0.7025	0.71
12.403	30.0	500	1.4914	0.6825	0.0625	0.4425	0.5175	0.6125	0.6675	0.6925	0.685	0.6875	0.685	0.6825	0.68	0.6825	0.69
12.403	30.96	516	1.4158	0.705	0.065	0.44	0.52	0.615	0.6725	0.685	0.7075	0.7	0.695	0.6925	0.6875	0.69	0.71
12.403	31.98	533	1.4566	0.7075	0.065	0.445	0.5275	0.6	0.665	0.66	0.6925	0.695	0.695	0.7025	0.7025	0.695	0.705
12.403	33.0	550	1.4925	0.6825	0.0625	0.45	0.5225	0.605	0.6775	0.6675	0.7025	0.6975	0.6975	0.6975	0.7025	0.6975	0.705
12.403	33.96	566	1.4863	0.6975	0.0675	0.4375	0.5325	0.61	0.6775	0.6775	0.6975	0.695	0.6975	0.69	0.6975	0.695	0.7075
12.403	34.98	583	1.4770	0.7075	0.0675	0.4625	0.525	0.625	0.6725	0.6775	0.695	0.6875	0.69	0.6925	0.695	0.705	0.7025
12.403	36.0	600	1.5347	0.6875	0.065	0.44	0.54	0.6175	0.6775	0.68	0.695	0.69	0.69	0.695	0.6875	0.6925	0.6975
12.403	36.96	616	1.5124	0.7025	0.065	0.445	0.5325	0.62	0.675	0.68	0.6975	0.7	0.6975	0.695	0.6975	0.7	0.71
12.403	37.98	633	1.5120	0.7075	0.065	0.4525	0.5375	0.6025	0.67	0.6925	0.6975	0.7025	0.695	0.6925	0.6975	0.6975	0.7075
12.403	39.0	650	1.5235	0.705	0.065	0.4575	0.5325	0.6175	0.675	0.6925	0.705	0.7	0.695	0.695	0.7	0.7025	0.72
12.403	39.96	666	1.5712	0.6875	0.065	0.46	0.5425	0.635	0.675	0.7	0.7025	0.7025	0.7	0.6975	0.6925	0.6925	0.6975
12.403	40.98	683	1.5346	0.705	0.065	0.455	0.545	0.6275	0.675	0.6975	0.6975	0.695	0.695	0.7025	0.7025	0.7025	0.71
12.403	42.0	700	1.5656	0.7025	0.065	0.47	0.55	0.635	0.6725	0.7	0.7075	0.7025	0.7025	0.7	0.7	0.6975	0.71
12.403	42.96	716	1.5712	0.695	0.065	0.465	0.5575	0.63	0.685	0.695	0.705	0.71	0.705	0.71	0.7025	0.705	0.7
12.403	43.98	733	1.5481	0.7075	0.0675	0.4675	0.545	0.63	0.6825	0.705	0.705	0.705	0.705	0.7	0.7025	0.705	0.715
12.403	45.0	750	1.5867	0.705	0.065	0.4625	0.5525	0.6325	0.68	0.6975	0.705	0.7025	0.7075	0.7025	0.6975	0.705	0.7075
12.403	45.96	766	1.5766	0.7075	0.0675	0.475	0.55	0.63	0.685	0.705	0.7025	0.7075	0.705	0.705	0.705	0.705	0.705
12.403	46.98	783	1.5702	0.71	0.065	0.475	0.5725	0.6475	0.6925	0.71	0.705	0.71	0.705	0.7025	0.7075	0.71	0.7225
12.403	48.0	800	1.5957	0.705	0.065	0.47	0.5575	0.6325	0.6875	0.715	0.71	0.7075	0.71	0.7025	0.7075	0.7025	0.7125
12.403	48.96	816	1.6135	0.705	0.065	0.4725	0.555	0.64	0.68	0.7	0.7075	0.71	0.705	0.7	0.6975	0.705	0.7125
12.403	49.98	833	1.5976	0.7075	0.065	0.4775	0.575	0.65	0.695	0.7025	0.7075	0.7075	0.7075	0.7075	0.7075	0.705	0.7125
12.403	51.0	850	1.5968	0.7175	0.065	0.4725	0.5675	0.6425	0.6875	0.7075	0.71	0.7125	0.7075	0.71	0.71	0.71	0.715
12.403	51.96	866	1.6005	0.705	0.065	0.475	0.565	0.6375	0.685	0.705	0.715	0.71	0.7025	0.7075	0.7075	0.705	0.715
12.403	52.98	883	1.6061	0.7125	0.065	0.48	0.57	0.6425	0.6925	0.71	0.715	0.71	0.705	0.71	0.7075	0.7075	0.7175
12.403	54.0	900	1.6100	0.7025	0.065	0.48	0.5675	0.6425	0.69	0.7075	0.7125	0.7075	0.705	0.7025	0.705	0.7025	0.7075
12.403	54.96	916	1.6125	0.705	0.065	0.48	0.57	0.6475	0.69	0.71	0.71	0.7075	0.705	0.705	0.7025	0.7025	0.71
12.403	55.98	933	1.6285	0.7	0.065	0.4825	0.57	0.6425	0.6875	0.705	0.71	0.705	0.7025	0.705	0.705	0.7	0.7025
12.403	57.0	950	1.6270	0.7025	0.065	0.48	0.57	0.6425	0.69	0.71	0.7175	0.705	0.71	0.7075	0.705	0.7025	0.7025
12.403	57.6	960	1.6250	0.7025	0.065	0.48	0.57	0.6425	0.6875	0.71	0.715	0.705	0.71	0.7075	0.7075	0.705	0.7075

Framework versions

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