layoutlmv3_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: 0.2675
• F1: 0.8608
• Precision: 0.8693
• Recall: 0.8526

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall
2.6697	0.9917	30	2.4757	0.0	0.0	0.0
1.2308	1.9835	60	1.2186	0.0	0.0	0.0
0.9265	2.9752	90	0.9238	0.4394	0.768	0.3077
0.6503	4.0	121	0.7466	0.4537	0.7254	0.3301
0.5476	4.9917	151	0.5594	0.5937	0.8650	0.4519
0.4409	5.9835	181	0.4283	0.6627	0.8615	0.5385
0.3424	6.9752	211	0.3609	0.7504	0.8531	0.6699
0.3238	8.0	242	0.3208	0.7973	0.8593	0.7436
0.2595	8.9917	272	0.2940	0.8215	0.8652	0.7821
0.2678	9.9835	302	0.2727	0.8405	0.9041	0.7853
0.2083	10.9752	332	0.2715	0.8508	0.9029	0.8045
0.198	12.0	363	0.2396	0.8495	0.8881	0.8141
0.1797	12.9917	393	0.2485	0.8454	0.8682	0.8237
0.2028	13.9835	423	0.2419	0.8424	0.8452	0.8397
0.1493	14.9752	453	0.2364	0.8668	0.9146	0.8237
0.1452	16.0	484	0.2462	0.8609	0.8904	0.8333
0.1538	16.9917	514	0.2632	0.8426	0.8360	0.8494
0.1111	17.9835	544	0.2555	0.8539	0.8553	0.8526
0.128	18.9752	574	0.2656	0.8475	0.8489	0.8462
0.0937	20.0	605	0.2675	0.8608	0.8693	0.8526

Framework versions

• Transformers 4.46.3
• Pytorch 2.10.0+cu128
• Datasets 3.1.0
• Tokenizers 0.20.3