layoutlm-custom

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1583
  • Noise: {'precision': 0.8818897637795275, 'recall': 0.8736349453978159, 'f1': 0.8777429467084641, 'number': 641}
  • Signal: {'precision': 0.861198738170347, 'recall': 0.853125, 'f1': 0.8571428571428572, 'number': 640}
  • Overall Precision: 0.8716
  • Overall Recall: 0.8634
  • Overall F1: 0.8675
  • Overall Accuracy: 0.9656

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: 3e-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
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Noise Signal Overall Precision Overall Recall Overall F1 Overall Accuracy
0.3882 1.0 18 0.2617 {'precision': 0.6654804270462633, 'recall': 0.5834633385335414, 'f1': 0.6217788861180383, 'number': 641} {'precision': 0.6149732620320856, 'recall': 0.5390625, 'f1': 0.5745212323064114, 'number': 640} 0.6402 0.5613 0.5982 0.8986
0.1694 2.0 36 0.1752 {'precision': 0.7387820512820513, 'recall': 0.719188767550702, 'f1': 0.7288537549407115, 'number': 641} {'precision': 0.709470304975923, 'recall': 0.690625, 'f1': 0.6999208234362629, 'number': 640} 0.7241 0.7049 0.7144 0.9296
0.1039 3.0 54 0.1356 {'precision': 0.7865168539325843, 'recall': 0.7644305772230889, 'f1': 0.7753164556962026, 'number': 641} {'precision': 0.77491961414791, 'recall': 0.753125, 'f1': 0.7638668779714739, 'number': 640} 0.7807 0.7588 0.7696 0.9439
0.064 4.0 72 0.1342 {'precision': 0.8220472440944881, 'recall': 0.8143525741029641, 'f1': 0.8181818181818181, 'number': 641} {'precision': 0.8028391167192429, 'recall': 0.7953125, 'f1': 0.7990580847723705, 'number': 640} 0.8125 0.8048 0.8086 0.9522
0.0433 5.0 90 0.1241 {'precision': 0.8544303797468354, 'recall': 0.8424336973478939, 'f1': 0.8483896307934014, 'number': 641} {'precision': 0.8320126782884311, 'recall': 0.8203125, 'f1': 0.8261211644374509, 'number': 640} 0.8432 0.8314 0.8373 0.9601
0.0293 6.0 108 0.1274 {'precision': 0.8650793650793651, 'recall': 0.8502340093603744, 'f1': 0.8575924468922109, 'number': 641} {'precision': 0.8378378378378378, 'recall': 0.8234375, 'f1': 0.830575256107171, 'number': 640} 0.8515 0.8368 0.8441 0.9617
0.0199 7.0 126 0.1372 {'precision': 0.8722397476340694, 'recall': 0.8627145085803433, 'f1': 0.8674509803921568, 'number': 641} {'precision': 0.8530805687203792, 'recall': 0.84375, 'f1': 0.8483896307934015, 'number': 640} 0.8627 0.8532 0.8579 0.9640
0.0139 8.0 144 0.1386 {'precision': 0.8839427662957074, 'recall': 0.8673946957878315, 'f1': 0.8755905511811023, 'number': 641} {'precision': 0.856687898089172, 'recall': 0.840625, 'f1': 0.8485804416403785, 'number': 640} 0.8703 0.8540 0.8621 0.9656
0.0126 9.0 162 0.1467 {'precision': 0.8829113924050633, 'recall': 0.8705148205928237, 'f1': 0.8766692851531814, 'number': 641} {'precision': 0.8541996830427893, 'recall': 0.8421875, 'f1': 0.848151062155783, 'number': 640} 0.8686 0.8564 0.8624 0.9654
0.0114 10.0 180 0.1531 {'precision': 0.8694968553459119, 'recall': 0.8627145085803433, 'f1': 0.8660924040720438, 'number': 641} {'precision': 0.8472440944881889, 'recall': 0.840625, 'f1': 0.8439215686274509, 'number': 640} 0.8584 0.8517 0.8550 0.9631
0.0099 11.0 198 0.1581 {'precision': 0.8703125, 'recall': 0.8689547581903276, 'f1': 0.8696330991412958, 'number': 641} {'precision': 0.8450704225352113, 'recall': 0.84375, 'f1': 0.8444096950742768, 'number': 640} 0.8577 0.8564 0.8570 0.9634
0.0064 12.0 216 0.1543 {'precision': 0.8866141732283465, 'recall': 0.8783151326053042, 'f1': 0.8824451410658307, 'number': 641} {'precision': 0.8643533123028391, 'recall': 0.85625, 'f1': 0.8602825745682888, 'number': 640} 0.8755 0.8673 0.8714 0.9659
0.0059 13.0 234 0.1628 {'precision': 0.8732394366197183, 'recall': 0.8705148205928237, 'f1': 0.871875, 'number': 641} {'precision': 0.8526645768025078, 'recall': 0.85, 'f1': 0.8513302034428795, 'number': 640} 0.8630 0.8603 0.8616 0.9645
0.0056 14.0 252 0.1587 {'precision': 0.878740157480315, 'recall': 0.8705148205928237, 'f1': 0.8746081504702194, 'number': 641} {'precision': 0.8580441640378549, 'recall': 0.85, 'f1': 0.8540031397174254, 'number': 640} 0.8684 0.8603 0.8643 0.9651
0.005 15.0 270 0.1583 {'precision': 0.8818897637795275, 'recall': 0.8736349453978159, 'f1': 0.8777429467084641, 'number': 641} {'precision': 0.861198738170347, 'recall': 0.853125, 'f1': 0.8571428571428572, 'number': 640} 0.8716 0.8634 0.8675 0.9656

Framework versions

  • Transformers 4.36.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0
Downloads last month
13
Safetensors
Model size
113M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.