2024-01-06_one_stage_subgraphs_weighted_txt_vis_conc_1_4_8_12_gate

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: 1.1285
• Accuracy: 0.7775
• Exit 0 Accuracy: 0.035
• Exit 1 Accuracy: 0.065
• Exit 2 Accuracy: 0.0625
• Exit 3 Accuracy: 0.0625
• Exit 4 Accuracy: 0.7775

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	2.6521	0.185	0.0625	0.0725	0.0625	0.0625	0.185
No log	1.98	33	2.5024	0.225	0.0625	0.0725	0.0625	0.0625	0.225
No log	3.0	50	2.3005	0.2825	0.0525	0.085	0.0625	0.0625	0.2825
No log	3.96	66	2.0909	0.38	0.0475	0.075	0.0625	0.0625	0.38
No log	4.98	83	1.8878	0.4675	0.045	0.065	0.0625	0.0625	0.4675
No log	6.0	100	1.6333	0.56	0.035	0.0625	0.0625	0.0625	0.56
No log	6.96	116	1.4491	0.625	0.035	0.0625	0.0625	0.0625	0.625
No log	7.98	133	1.2641	0.6625	0.03	0.0625	0.0625	0.0625	0.6625
No log	9.0	150	1.1697	0.6875	0.035	0.0625	0.0625	0.0625	0.6875
No log	9.96	166	1.0498	0.7275	0.03	0.0625	0.0625	0.0625	0.7275
No log	10.98	183	1.0408	0.715	0.03	0.0625	0.0625	0.0625	0.715
No log	12.0	200	0.9298	0.7575	0.04	0.065	0.0625	0.0625	0.7575
No log	12.96	216	0.9264	0.7425	0.0375	0.065	0.0625	0.0625	0.7425
No log	13.98	233	0.9073	0.7475	0.0375	0.065	0.0625	0.0625	0.7475
No log	15.0	250	0.8787	0.7525	0.0375	0.065	0.0625	0.0625	0.7525
No log	15.96	266	0.9141	0.755	0.0375	0.06	0.0625	0.0625	0.755
No log	16.98	283	0.9292	0.7475	0.0375	0.0525	0.0625	0.0625	0.7475
No log	18.0	300	0.9011	0.7725	0.0375	0.0475	0.0625	0.0625	0.7725
No log	18.96	316	0.9439	0.765	0.0375	0.05	0.0625	0.0625	0.765
No log	19.98	333	0.9692	0.7675	0.04	0.0575	0.0625	0.0625	0.7675
No log	21.0	350	0.9501	0.775	0.035	0.065	0.0625	0.0625	0.775
No log	21.96	366	0.9702	0.7725	0.035	0.06	0.0625	0.0625	0.7725
No log	22.98	383	1.0124	0.7675	0.0425	0.06	0.0625	0.0625	0.7675
No log	24.0	400	0.9772	0.78	0.0375	0.06	0.0625	0.0625	0.78
No log	24.96	416	1.0035	0.7725	0.0375	0.06	0.0625	0.0625	0.7725
No log	25.98	433	1.0374	0.77	0.0375	0.06	0.0625	0.0625	0.77
No log	27.0	450	1.0284	0.7675	0.0375	0.0625	0.0625	0.0625	0.7675
No log	27.96	466	1.0110	0.7825	0.04	0.0625	0.0625	0.0625	0.7825
No log	28.98	483	1.0291	0.7725	0.04	0.0625	0.0625	0.0625	0.7725
0.3746	30.0	500	1.0285	0.775	0.04	0.0625	0.0625	0.0625	0.775
0.3746	30.96	516	1.0399	0.7775	0.04	0.0625	0.0625	0.0625	0.7775
0.3746	31.98	533	1.0550	0.78	0.04	0.0625	0.0625	0.0625	0.78
0.3746	33.0	550	1.0606	0.7775	0.04	0.065	0.0625	0.0625	0.7775
0.3746	33.96	566	1.0618	0.7775	0.04	0.065	0.0625	0.0625	0.7775
0.3746	34.98	583	1.0651	0.7825	0.04	0.065	0.0625	0.0625	0.7825
0.3746	36.0	600	1.0646	0.775	0.0375	0.065	0.0625	0.0625	0.775
0.3746	36.96	616	1.0777	0.775	0.0375	0.065	0.0625	0.0625	0.775
0.3746	37.98	633	1.0896	0.7725	0.04	0.065	0.0625	0.0625	0.7725
0.3746	39.0	650	1.0834	0.785	0.04	0.065	0.0625	0.0625	0.785
0.3746	39.96	666	1.0875	0.78	0.0375	0.065	0.0625	0.0625	0.78
0.3746	40.98	683	1.0831	0.7825	0.0375	0.065	0.0625	0.0625	0.7825
0.3746	42.0	700	1.0980	0.78	0.0375	0.065	0.0625	0.0625	0.78
0.3746	42.96	716	1.1054	0.7725	0.0375	0.065	0.0625	0.0625	0.7725
0.3746	43.98	733	1.1083	0.7725	0.035	0.065	0.0625	0.0625	0.7725
0.3746	45.0	750	1.1166	0.775	0.035	0.065	0.0625	0.0625	0.775
0.3746	45.96	766	1.1093	0.775	0.035	0.065	0.0625	0.0625	0.775
0.3746	46.98	783	1.1114	0.7725	0.035	0.065	0.0625	0.0625	0.7725
0.3746	48.0	800	1.1183	0.7725	0.035	0.065	0.0625	0.0625	0.7725
0.3746	48.96	816	1.1178	0.7725	0.035	0.065	0.0625	0.0625	0.7725
0.3746	49.98	833	1.1169	0.775	0.035	0.065	0.0625	0.0625	0.775
0.3746	51.0	850	1.1187	0.7725	0.035	0.065	0.0625	0.0625	0.7725
0.3746	51.96	866	1.1183	0.7775	0.035	0.065	0.0625	0.0625	0.7775
0.3746	52.98	883	1.1227	0.775	0.035	0.065	0.0625	0.0625	0.775
0.3746	54.0	900	1.1253	0.78	0.035	0.065	0.0625	0.0625	0.78
0.3746	54.96	916	1.1269	0.7775	0.035	0.065	0.0625	0.0625	0.7775
0.3746	55.98	933	1.1279	0.7775	0.035	0.065	0.0625	0.0625	0.7775
0.3746	57.0	950	1.1283	0.7775	0.035	0.065	0.0625	0.0625	0.7775
0.3746	57.6	960	1.1285	0.7775	0.035	0.065	0.0625	0.0625	0.7775

Framework versions

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