Bunesh-VirtuonAI/layoutlmv3-finetuned-cord_100

layoutlmv3-finetuned-funsd_100

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

• Loss: 0.5728
• Precision: 0.8172
• Recall: 0.8664
• F1: 0.8411
• Accuracy: 0.8318

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: 250

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.83	25	1.3530	0.2996	0.3040	0.3018	0.5402
No log	1.67	50	0.9373	0.6537	0.7193	0.6850	0.7412
No log	2.5	75	0.7492	0.7574	0.8018	0.7790	0.7748
No log	3.33	100	0.6587	0.7721	0.8097	0.7905	0.7900
No log	4.17	125	0.6224	0.7808	0.8336	0.8063	0.8005
No log	5.0	150	0.5720	0.7870	0.8445	0.8148	0.8171
No log	5.83	175	0.5343	0.8164	0.8549	0.8352	0.8250
No log	6.67	200	0.5856	0.8139	0.8604	0.8365	0.8268
No log	7.5	225	0.5787	0.8166	0.8624	0.8388	0.8266
No log	8.33	250	0.5728	0.8172	0.8664	0.8411	0.8318

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.2
• Datasets 2.18.0
• Tokenizers 0.15.2