Layoutlmv3InvoiceCzechV013

This model is a fine-tuned version of TomasFAV/Layoutlmv3InvoiceCzechV01 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0405
• Precision: 0.9167
• Recall: 0.9306
• F1: 0.9236
• Accuracy: 0.9925

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: 8
• eval_batch_size: 1
• 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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	23	0.0980	0.7728	0.7022	0.7358	0.9754
No log	2.0	46	0.0724	0.7531	0.8308	0.7900	0.9796
No log	3.0	69	0.0544	0.8523	0.8494	0.8508	0.9877
No log	4.0	92	0.0465	0.8307	0.9052	0.8664	0.9881
No log	5.0	115	0.0447	0.8613	0.9036	0.8819	0.9896
No log	6.0	138	0.0478	0.8941	0.9002	0.8971	0.9907
No log	7.0	161	0.0400	0.8911	0.9137	0.9023	0.9911
No log	8.0	184	0.0409	0.9064	0.9171	0.9117	0.9925
No log	9.0	207	0.0410	0.9037	0.9205	0.9120	0.9919
No log	10.0	230	0.0432	0.8805	0.9222	0.9008	0.9910
No log	11.0	253	0.0396	0.9039	0.9391	0.9212	0.9926
No log	12.0	276	0.0406	0.9128	0.9205	0.9166	0.9923
No log	13.0	299	0.0380	0.9117	0.9255	0.9186	0.9927
No log	14.0	322	0.0391	0.9064	0.9340	0.9200	0.9926
No log	15.0	345	0.0393	0.9066	0.9357	0.9209	0.9926
No log	16.0	368	0.0416	0.9176	0.9239	0.9207	0.9924
No log	17.0	391	0.0382	0.9097	0.9374	0.9233	0.9928
No log	18.0	414	0.0405	0.9183	0.9323	0.9253	0.9926
No log	19.0	437	0.0402	0.9165	0.9289	0.9227	0.9927
No log	20.0	460	0.0398	0.9147	0.9255	0.9201	0.9925

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2