EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-08-12_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: 1.2733
• Accuracy: 0.7825
• Exit 0 Accuracy: 0.0775

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
No log	0.96	16	2.6800	0.13	0.0675
No log	1.98	33	2.4446	0.28	0.0725
No log	3.0	50	2.1924	0.37	0.0675
No log	3.96	66	1.8733	0.5175	0.07
No log	4.98	83	1.6056	0.6075	0.0775
No log	6.0	100	1.3480	0.6725	0.08
No log	6.96	116	1.1393	0.735	0.07
No log	7.98	133	1.0738	0.7375	0.07
No log	9.0	150	0.9271	0.7725	0.075
No log	9.96	166	0.8885	0.7675	0.085
No log	10.98	183	0.8669	0.76	0.075
No log	12.0	200	0.8547	0.7775	0.0725
No log	12.96	216	0.8633	0.76	0.07
No log	13.98	233	0.8498	0.7675	0.075
No log	15.0	250	0.9608	0.7675	0.0675
No log	15.96	266	0.8952	0.7875	0.08
No log	16.98	283	0.9486	0.7575	0.0725
No log	18.0	300	0.9826	0.765	0.0825
No log	18.96	316	1.0230	0.7625	0.09
No log	19.98	333	1.0961	0.76	0.0875
No log	21.0	350	1.0083	0.785	0.07
No log	21.96	366	1.0394	0.7725	0.0725
No log	22.98	383	1.0825	0.78	0.085
No log	24.0	400	1.0789	0.77	0.075
No log	24.96	416	1.1030	0.7725	0.0925
No log	25.98	433	1.1252	0.775	0.075
No log	27.0	450	1.1333	0.7725	0.0725
No log	27.96	466	1.1416	0.765	0.0775
No log	28.98	483	1.1442	0.7775	0.0775
1.6768	30.0	500	1.1620	0.7825	0.1025
1.6768	30.96	516	1.1617	0.7825	0.0775
1.6768	31.98	533	1.1788	0.775	0.0875
1.6768	33.0	550	1.1858	0.7725	0.0825
1.6768	33.96	566	1.1842	0.7825	0.0725
1.6768	34.98	583	1.1964	0.785	0.085
1.6768	36.0	600	1.2034	0.78	0.075
1.6768	36.96	616	1.2050	0.7825	0.07
1.6768	37.98	633	1.2111	0.7825	0.075
1.6768	39.0	650	1.2217	0.785	0.0925
1.6768	39.96	666	1.2510	0.7775	0.105
1.6768	40.98	683	1.2512	0.7825	0.0825
1.6768	42.0	700	1.2529	0.7775	0.0775
1.6768	42.96	716	1.2557	0.78	0.0725
1.6768	43.98	733	1.2615	0.7775	0.0775
1.6768	45.0	750	1.2621	0.78	0.0825
1.6768	45.96	766	1.2613	0.785	0.075
1.6768	46.98	783	1.2614	0.78	0.075
1.6768	48.0	800	1.2598	0.7825	0.075
1.6768	48.96	816	1.2650	0.7825	0.085
1.6768	49.98	833	1.2665	0.7825	0.08
1.6768	51.0	850	1.2673	0.785	0.0775
1.6768	51.96	866	1.2626	0.7775	0.075
1.6768	52.98	883	1.2643	0.7825	0.075
1.6768	54.0	900	1.2702	0.78	0.0775
1.6768	54.96	916	1.2723	0.78	0.0775
1.6768	55.98	933	1.2730	0.7825	0.0775
1.6768	57.0	950	1.2732	0.7825	0.0775
1.6768	57.6	960	1.2733	0.7825	0.0775

Framework versions

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