layoutlmv3-ap8_3_stitched_32_batch_4e5_2k

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: 0.3092
• Precision: 0.4348
• Recall: 0.1205
• F1: 0.1887
• Accuracy: 0.9741

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: 4e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• 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
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	62.5	250	0.2442	0.3043	0.0843	0.1321	0.9731
0.237	125.0	500	0.2593	0.3182	0.1687	0.2205	0.9731
0.237	187.5	750	0.2878	0.4167	0.1205	0.1869	0.9734
0.0016	250.0	1000	0.2873	0.4242	0.1687	0.2414	0.9736
0.0016	312.5	1250	0.3024	0.4286	0.1084	0.1731	0.9734
0.0006	375.0	1500	0.3034	0.4286	0.1084	0.1731	0.9739
0.0006	437.5	1750	0.3067	0.4074	0.1325	0.2	0.9737
0.0004	500.0	2000	0.3092	0.4348	0.1205	0.1887	0.9741

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Tokenizers 0.21.1