layoutlmv3-for-complete-receipt-understanding

This model is a fine-tuned version of NLPmonster/layoutlmv3-for-receipt-understanding on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4421
• Precision: 0.6502
• Recall: 0.6607
• F1: 0.6554
• Accuracy: 0.8760

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: 5e-05
• train_batch_size: 5
• eval_batch_size: 5
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.2604	0.4673	50	0.6573	0.5077	0.4747	0.4906	0.7863
0.5256	0.9346	100	0.4824	0.5732	0.6521	0.6101	0.8465
0.4529	1.4019	150	0.4461	0.5926	0.6509	0.6204	0.8464
0.3767	1.8692	200	0.4265	0.5747	0.6740	0.6204	0.8492
0.3403	2.3364	250	0.4557	0.5564	0.6083	0.5812	0.8451
0.331	2.8037	300	0.4065	0.6384	0.6671	0.6524	0.8669
0.2984	3.2710	350	0.3820	0.6411	0.6411	0.6411	0.8729
0.2763	3.7383	400	0.4078	0.6104	0.6037	0.6070	0.8576
0.2626	4.2056	450	0.4203	0.6268	0.6164	0.6216	0.8589
0.2456	4.6729	500	0.3960	0.6240	0.6406	0.6322	0.8686
0.2078	5.1402	550	0.4074	0.6401	0.6290	0.6345	0.8709
0.1859	5.6075	600	0.3853	0.6511	0.6601	0.6556	0.8733
0.2059	6.0748	650	0.3845	0.6539	0.6509	0.6524	0.8772
0.1649	6.5421	700	0.4128	0.6298	0.6486	0.6390	0.8706
0.1599	7.0093	750	0.4328	0.6302	0.6578	0.6437	0.8644
0.1437	7.4766	800	0.4100	0.6510	0.6469	0.6489	0.8727
0.1377	7.9439	850	0.4409	0.6317	0.6699	0.6503	0.8711
0.1164	8.4112	900	0.4331	0.6301	0.6642	0.6467	0.8717
0.1149	8.8785	950	0.4523	0.6466	0.6555	0.6510	0.8712
0.1096	9.3458	1000	0.4421	0.6502	0.6607	0.6554	0.8760

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1