project-ocr

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

• Loss: 0.9877
• Precision: 0.7516
• Recall: 0.8039
• F1: 0.7769
• Accuracy: 0.8103

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: 5
• eval_batch_size: 5
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.83	50	2.6184	0.4355	0.5404	0.4823	0.4338
No log	1.67	100	1.8766	0.5912	0.6018	0.5964	0.5620
No log	2.5	150	1.6165	0.5737	0.6347	0.6027	0.6150
No log	3.33	200	1.4317	0.5732	0.6737	0.6194	0.6944
No log	4.17	250	1.2787	0.6190	0.7126	0.6625	0.7347
No log	5.0	300	1.1632	0.6729	0.7560	0.7120	0.7759
No log	5.83	350	1.0990	0.6980	0.7665	0.7306	0.7857
No log	6.67	400	1.0327	0.7125	0.7792	0.7444	0.7946
No log	7.5	450	0.9994	0.7526	0.8016	0.7764	0.8065
1.6589	8.33	500	0.9877	0.7516	0.8039	0.7769	0.8103

Framework versions

• Transformers 4.27.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2