layoutlmv3-finetuned-funsd

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

• Loss: 1.0068
• Precision: 0.7998
• Recall: 0.8376
• F1: 0.8182
• Accuracy: 0.8261

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 2500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	3.33	250	0.5828	0.7015	0.8033	0.7490	0.8022
0.6702	6.67	500	0.5765	0.7499	0.8073	0.7775	0.8253
0.6702	10.0	750	0.7082	0.7755	0.8236	0.7988	0.8160
0.1797	13.33	1000	0.7819	0.7807	0.8366	0.8077	0.8256
0.1797	16.67	1250	0.8199	0.7997	0.8311	0.8151	0.8227
0.0745	20.0	1500	0.9025	0.7943	0.8286	0.8111	0.8231
0.0745	23.33	1750	0.9159	0.7941	0.8470	0.8197	0.8248
0.041	26.67	2000	1.0012	0.7989	0.8385	0.8182	0.8210
0.041	30.0	2250	0.9852	0.8024	0.8450	0.8231	0.8301
0.0246	33.33	2500	1.0068	0.7998	0.8376	0.8182	0.8261

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3