layoutlmv3-finetuned-full

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

• Loss: 0.0613
• Precision: 0.9339
• Recall: 0.9517
• F1: 0.9427
• Accuracy: 0.9888

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: 3
• total_train_batch_size: 6
• 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
• training_steps: 2500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.5201	250	0.3041	0.4864	0.5643	0.5225	0.9219
0.4848	1.0416	500	0.1620	0.7495	0.8031	0.7753	0.9652
0.4848	1.5617	750	0.1195	0.8386	0.8662	0.8522	0.9745
0.1555	2.0832	1000	0.0996	0.8764	0.9025	0.8892	0.9790
0.1555	2.6033	1250	0.0765	0.8984	0.9285	0.9132	0.9828
0.0941	3.1248	1500	0.0662	0.9207	0.9387	0.9296	0.9864
0.0941	3.6449	1750	0.0658	0.9361	0.9452	0.9406	0.9875
0.0643	4.1664	2000	0.0630	0.9317	0.9508	0.9411	0.9886
0.0643	4.6865	2250	0.0589	0.9338	0.9503	0.9420	0.9892
0.0503	5.2080	2500	0.0613	0.9339	0.9517	0.9427	0.9888

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0