docintel-extractor

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: 1.4371
• Precision: 0.8960
• Recall: 0.9042
• F1: 0.9001
• Accuracy: 0.8642

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	19	2.3352	0.5368	0.4916	0.5132	0.5602
No log	2.0	38	1.2024	0.7210	0.8358	0.7742	0.7811
No log	3.0	57	1.0774	0.8171	0.8784	0.8466	0.8039
No log	4.0	76	1.0967	0.8018	0.8530	0.8266	0.8176
No log	5.0	95	1.1609	0.8495	0.8555	0.8525	0.8360
No log	6.0	114	1.0465	0.8610	0.9133	0.8864	0.8575
No log	7.0	133	1.1293	0.8822	0.9072	0.8946	0.8493
No log	8.0	152	1.3029	0.8352	0.8525	0.8437	0.8229
No log	9.0	171	1.2912	0.8836	0.8925	0.8880	0.8538
No log	10.0	190	1.3468	0.8949	0.8976	0.8963	0.8371
No log	11.0	209	1.4046	0.8896	0.8905	0.8901	0.8442
No log	12.0	228	1.3960	0.9076	0.8915	0.8995	0.8619
No log	13.0	247	1.4927	0.8991	0.8941	0.8966	0.8618
No log	14.0	266	1.4369	0.8960	0.9042	0.9001	0.8642
No log	15.0	285	1.4735	0.8989	0.8971	0.8980	0.8617

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.8.5
• Tokenizers 0.22.2