layoutlmv3-finetuned-invoice

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: 0.8510
• Precision: 0.9058
• Recall: 0.9175
• F1: 0.9116
• Accuracy: 0.8556

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: 1e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• training_steps: 2000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.3333	100	0.6411	0.7779	0.8316	0.8038	0.7951
No log	2.6667	200	0.5235	0.8209	0.8629	0.8414	0.8238
No log	4.0	300	0.5130	0.8738	0.9081	0.8906	0.8504
No log	5.3333	400	0.5742	0.8848	0.9081	0.8963	0.8431
0.5233	6.6667	500	0.6276	0.8610	0.8927	0.8766	0.8374
0.5233	8.0	600	0.6887	0.8818	0.9041	0.8928	0.8357
0.5233	9.3333	700	0.6323	0.8930	0.9165	0.9046	0.8628
0.5233	10.6667	800	0.6644	0.8878	0.9195	0.9034	0.8538
0.5233	12.0	900	0.7365	0.9138	0.9210	0.9174	0.8580
0.1181	13.3333	1000	0.7774	0.8939	0.9210	0.9073	0.8549
0.1181	14.6667	1100	0.8265	0.9090	0.9175	0.9132	0.8557
0.1181	16.0	1200	0.8112	0.9023	0.9265	0.9142	0.8546
0.1181	17.3333	1300	0.8212	0.9075	0.9160	0.9117	0.8596
0.1181	18.6667	1400	0.8931	0.8999	0.9151	0.9074	0.8509
0.0443	20.0	1500	0.8510	0.9058	0.9175	0.9116	0.8556
0.0443	21.3333	1600	0.8318	0.9016	0.9235	0.9124	0.8612
0.0443	22.6667	1700	0.8783	0.9065	0.9146	0.9105	0.8519
0.0443	24.0	1800	0.8964	0.9023	0.9126	0.9074	0.8527
0.0443	25.3333	1900	0.8890	0.9054	0.9175	0.9114	0.8580
0.0205	26.6667	2000	0.8992	0.9004	0.9165	0.9084	0.8552

Framework versions

• Transformers 4.46.0.dev0
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.20.1