layoutlmv3-finetuned-invoice_ConControl_v3

This model is a fine-tuned version of microsoft/layoutlmv3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.3493
• Precision: 0.0049
• Recall: 0.0587
• F1: 0.0090
• Accuracy: 0.0569

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 1500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	2	2.6128	0.0036	0.0475	0.0067	0.0087
No log	2.0	4	2.6071	0.0036	0.0475	0.0067	0.0087
No log	3.0	6	2.5970	0.0034	0.0447	0.0063	0.0087
No log	4.0	8	2.5823	0.0036	0.0475	0.0067	0.0089
No log	5.0	10	2.5632	0.0041	0.0531	0.0075	0.0092
No log	6.0	12	2.5395	0.0045	0.0587	0.0084	0.0100
No log	7.0	14	2.5113	0.0045	0.0587	0.0084	0.0110
No log	8.0	16	2.4784	0.0046	0.0587	0.0085	0.0135
No log	9.0	18	2.4407	0.0044	0.0559	0.0082	0.0189
No log	10.0	20	2.3978	0.0045	0.0559	0.0083	0.0296
No log	11.0	22	2.3493	0.0049	0.0587	0.0090	0.0569

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0+cu118
• Datasets 3.4.1
• Tokenizers 0.21.1