lilt-en-funsd

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

• eval_loss: 1.9291
• eval_ANSWER: {'precision': 0.0166358595194085, 'recall': 0.044063647490820076, 'f1': 0.024152968802415294, 'number': 817}
• eval_HEADER: {'precision': 0.004098360655737705, 'recall': 0.008403361344537815, 'f1': 0.005509641873278237, 'number': 119}
• eval_QUESTION: {'precision': 0.08307501549907005, 'recall': 0.2488393686165274, 'f1': 0.12456425749477108, 'number': 1077}
• eval_overall_precision: 0.0541
• eval_overall_recall: 0.1515
• eval_overall_f1: 0.0798
• eval_overall_accuracy: 0.1625
• eval_runtime: 4.0045
• eval_samples_per_second: 12.486
• eval_steps_per_second: 1.748
• step: 0

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 2500

Framework versions

• Transformers 4.34.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1