EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-08-26_txt_vis_concat_enc_5_6_7_8_ramp

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: 1.7273
• Accuracy: 0.725
• Exit 0 Accuracy: 0.07
• Exit 1 Accuracy: 0.7175
• Exit 2 Accuracy: 0.725
• Exit 3 Accuracy: 0.72
• Exit 4 Accuracy: 0.72

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	2.6945	0.1475	0.0675	0.0625	0.0625	0.0625	0.0625
No log	1.98	33	2.5294	0.235	0.0725	0.1775	0.0625	0.0575	0.0625
No log	3.0	50	2.3532	0.2975	0.0575	0.1925	0.155	0.1475	0.0625
No log	3.96	66	2.1334	0.3675	0.0675	0.21	0.2225	0.145	0.0625
No log	4.98	83	1.8627	0.5075	0.0625	0.245	0.245	0.2075	0.0625
No log	6.0	100	1.6181	0.575	0.0625	0.2875	0.2975	0.2875	0.0625
No log	6.96	116	1.4648	0.5825	0.0675	0.335	0.2925	0.3225	0.08
No log	7.98	133	1.3335	0.6525	0.0775	0.43	0.355	0.37	0.175
No log	9.0	150	1.2244	0.6625	0.0725	0.395	0.4025	0.44	0.24
No log	9.96	166	1.0634	0.7325	0.07	0.4425	0.415	0.5025	0.295
No log	10.98	183	1.0834	0.6975	0.07	0.4375	0.435	0.5075	0.3375
No log	12.0	200	0.9979	0.7325	0.0675	0.495	0.5025	0.54	0.445
No log	12.96	216	1.0157	0.7125	0.0625	0.475	0.54	0.575	0.4475
No log	13.98	233	1.0106	0.715	0.06	0.51	0.555	0.595	0.535
No log	15.0	250	1.0085	0.73	0.0575	0.555	0.6275	0.61	0.575
No log	15.96	266	1.0560	0.72	0.06	0.5775	0.6175	0.5975	0.6275
No log	16.98	283	1.0970	0.7275	0.055	0.61	0.6475	0.6525	0.675
No log	18.0	300	1.1175	0.7425	0.055	0.6025	0.6575	0.65	0.6725
No log	18.96	316	1.1457	0.7225	0.055	0.635	0.6625	0.6575	0.6775
No log	19.98	333	1.1841	0.7325	0.0625	0.6475	0.66	0.66	0.71
No log	21.0	350	1.2550	0.72	0.055	0.6675	0.6725	0.665	0.6925
No log	21.96	366	1.1865	0.7375	0.0525	0.6575	0.6675	0.665	0.71
No log	22.98	383	1.2642	0.7375	0.0525	0.6625	0.7	0.6775	0.71
No log	24.0	400	1.2972	0.7375	0.0525	0.675	0.6875	0.6775	0.71
No log	24.96	416	1.3043	0.7375	0.055	0.68	0.7075	0.695	0.7125
No log	25.98	433	1.3557	0.73	0.0575	0.675	0.7075	0.705	0.7175
No log	27.0	450	1.3476	0.7525	0.0575	0.6875	0.705	0.705	0.7325
No log	27.96	466	1.4352	0.7275	0.0575	0.7	0.7275	0.7175	0.73
No log	28.98	483	1.4874	0.7175	0.055	0.695	0.7	0.71	0.715
1.4115	30.0	500	1.4958	0.715	0.055	0.7025	0.6975	0.7175	0.725
1.4115	30.96	516	1.5134	0.7225	0.0575	0.72	0.735	0.73	0.7425
1.4115	31.98	533	1.5730	0.72	0.0575	0.7175	0.715	0.73	0.7325
1.4115	33.0	550	1.5390	0.7225	0.0575	0.7325	0.725	0.72	0.7275
1.4115	33.96	566	1.5308	0.73	0.0575	0.72	0.7125	0.7125	0.7225
1.4115	34.98	583	1.5720	0.7275	0.0575	0.73	0.7175	0.7175	0.73
1.4115	36.0	600	1.6400	0.7225	0.06	0.715	0.715	0.715	0.7275
1.4115	36.96	616	1.5729	0.745	0.06	0.72	0.73	0.73	0.7325
1.4115	37.98	633	1.5931	0.7375	0.065	0.73	0.7325	0.735	0.7325
1.4115	39.0	650	1.6346	0.725	0.065	0.7225	0.7175	0.73	0.7325
1.4115	39.96	666	1.6544	0.7225	0.065	0.72	0.7225	0.72	0.7225
1.4115	40.98	683	1.6793	0.725	0.065	0.735	0.7225	0.7325	0.7325
1.4115	42.0	700	1.6609	0.73	0.065	0.715	0.7225	0.73	0.7275
1.4115	42.96	716	1.6950	0.7175	0.0675	0.72	0.7275	0.7225	0.7225
1.4115	43.98	733	1.7207	0.7275	0.07	0.725	0.7275	0.7325	0.7275
1.4115	45.0	750	1.6978	0.7325	0.07	0.725	0.7225	0.7225	0.725
1.4115	45.96	766	1.6981	0.7325	0.07	0.725	0.735	0.7275	0.7275
1.4115	46.98	783	1.7356	0.73	0.07	0.7075	0.72	0.715	0.715
1.4115	48.0	800	1.7017	0.73	0.07	0.7175	0.7225	0.7275	0.725
1.4115	48.96	816	1.7040	0.73	0.07	0.72	0.725	0.7225	0.7275
1.4115	49.98	833	1.7298	0.7225	0.07	0.715	0.73	0.73	0.7225
1.4115	51.0	850	1.7115	0.725	0.07	0.72	0.725	0.725	0.7225
1.4115	51.96	866	1.7245	0.7325	0.07	0.71	0.7275	0.72	0.7225
1.4115	52.98	883	1.6998	0.74	0.07	0.72	0.725	0.725	0.7225
1.4115	54.0	900	1.7205	0.7325	0.07	0.7175	0.7275	0.725	0.7225
1.4115	54.96	916	1.7318	0.7275	0.07	0.72	0.73	0.7225	0.7225
1.4115	55.98	933	1.7341	0.7275	0.07	0.7175	0.7275	0.72	0.7225
1.4115	57.0	950	1.7280	0.7225	0.07	0.7175	0.725	0.7175	0.72
1.4115	57.6	960	1.7273	0.725	0.07	0.7175	0.725	0.72	0.72

Framework versions

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