lilt-en-funsd

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

  • Loss: 1.6681
  • Answer: {'precision': 0.8778173190984578, 'recall': 0.9057527539779682, 'f1': 0.891566265060241, 'number': 817}
  • Header: {'precision': 0.6036036036036037, 'recall': 0.5630252100840336, 'f1': 0.582608695652174, 'number': 119}
  • Question: {'precision': 0.9024839006439742, 'recall': 0.9108635097493036, 'f1': 0.9066543438077633, 'number': 1077}
  • Overall Precision: 0.8760
  • Overall Recall: 0.8882
  • Overall F1: 0.8821
  • Overall Accuracy: 0.8030

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
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Answer Header Question Overall Precision Overall Recall Overall F1 Overall Accuracy
0.431 10.5263 200 1.0598 {'precision': 0.8017718715393134, 'recall': 0.8861689106487148, 'f1': 0.841860465116279, 'number': 817} {'precision': 0.4228187919463087, 'recall': 0.5294117647058824, 'f1': 0.47014925373134325, 'number': 119} {'precision': 0.8755935422602089, 'recall': 0.8560817084493965, 'f1': 0.8657276995305164, 'number': 1077} 0.8119 0.8490 0.8300 0.7774
0.0542 21.0526 400 1.2173 {'precision': 0.8382352941176471, 'recall': 0.9069767441860465, 'f1': 0.8712522045855379, 'number': 817} {'precision': 0.5350877192982456, 'recall': 0.5126050420168067, 'f1': 0.5236051502145922, 'number': 119} {'precision': 0.8882521489971347, 'recall': 0.8635097493036211, 'f1': 0.8757062146892656, 'number': 1077} 0.8469 0.8604 0.8536 0.8016
0.014 31.5789 600 1.2955 {'precision': 0.8415051311288484, 'recall': 0.9033047735618115, 'f1': 0.8713105076741442, 'number': 817} {'precision': 0.6210526315789474, 'recall': 0.4957983193277311, 'f1': 0.5514018691588785, 'number': 119} {'precision': 0.8972477064220183, 'recall': 0.9080779944289693, 'f1': 0.9026303645592985, 'number': 1077} 0.8608 0.8818 0.8712 0.8160
0.0064 42.1053 800 1.2848 {'precision': 0.8696186961869619, 'recall': 0.8653610771113831, 'f1': 0.8674846625766871, 'number': 817} {'precision': 0.5193798449612403, 'recall': 0.5630252100840336, 'f1': 0.5403225806451614, 'number': 119} {'precision': 0.858274647887324, 'recall': 0.9052924791086351, 'f1': 0.8811568007230005, 'number': 1077} 0.8417 0.8689 0.8550 0.8222
0.0037 52.6316 1000 1.5983 {'precision': 0.8530751708428246, 'recall': 0.9167686658506732, 'f1': 0.8837758112094395, 'number': 817} {'precision': 0.5658914728682171, 'recall': 0.6134453781512605, 'f1': 0.5887096774193549, 'number': 119} {'precision': 0.8946360153256705, 'recall': 0.8672237697307336, 'f1': 0.8807166430928807, 'number': 1077} 0.8562 0.8723 0.8642 0.7916
0.0034 63.1579 1200 1.5936 {'precision': 0.85, 'recall': 0.9155446756425949, 'f1': 0.881555686505598, 'number': 817} {'precision': 0.5619047619047619, 'recall': 0.4957983193277311, 'f1': 0.5267857142857143, 'number': 119} {'precision': 0.8912442396313364, 'recall': 0.8978644382544104, 'f1': 0.8945420906567992, 'number': 1077} 0.8570 0.8813 0.8690 0.8102
0.0021 73.6842 1400 1.4765 {'precision': 0.8558139534883721, 'recall': 0.9008567931456548, 'f1': 0.877757901013715, 'number': 817} {'precision': 0.5619047619047619, 'recall': 0.4957983193277311, 'f1': 0.5267857142857143, 'number': 119} {'precision': 0.885036496350365, 'recall': 0.9006499535747446, 'f1': 0.892774965485504, 'number': 1077} 0.8564 0.8768 0.8665 0.8010
0.0009 84.2105 1600 1.6681 {'precision': 0.8778173190984578, 'recall': 0.9057527539779682, 'f1': 0.891566265060241, 'number': 817} {'precision': 0.6036036036036037, 'recall': 0.5630252100840336, 'f1': 0.582608695652174, 'number': 119} {'precision': 0.9024839006439742, 'recall': 0.9108635097493036, 'f1': 0.9066543438077633, 'number': 1077} 0.8760 0.8882 0.8821 0.8030
0.0003 94.7368 1800 1.6379 {'precision': 0.8595238095238096, 'recall': 0.8837209302325582, 'f1': 0.8714544357272178, 'number': 817} {'precision': 0.5929203539823009, 'recall': 0.5630252100840336, 'f1': 0.5775862068965517, 'number': 119} {'precision': 0.896709323583181, 'recall': 0.9108635097493036, 'f1': 0.9037309995393827, 'number': 1077} 0.8647 0.8793 0.8719 0.7986
0.0002 105.2632 2000 1.7186 {'precision': 0.8644859813084113, 'recall': 0.9057527539779682, 'f1': 0.8846383741781233, 'number': 817} {'precision': 0.5675675675675675, 'recall': 0.5294117647058824, 'f1': 0.5478260869565218, 'number': 119} {'precision': 0.8921658986175115, 'recall': 0.8987929433611885, 'f1': 0.8954671600370029, 'number': 1077} 0.8631 0.8798 0.8713 0.7978
0.0003 115.7895 2200 1.6765 {'precision': 0.8690476190476191, 'recall': 0.8935128518971848, 'f1': 0.8811104405552203, 'number': 817} {'precision': 0.5726495726495726, 'recall': 0.5630252100840336, 'f1': 0.5677966101694915, 'number': 119} {'precision': 0.8934802571166207, 'recall': 0.903435468895079, 'f1': 0.8984302862419206, 'number': 1077} 0.8651 0.8793 0.8721 0.8000
0.0003 126.3158 2400 1.7309 {'precision': 0.8817852834740652, 'recall': 0.8947368421052632, 'f1': 0.888213851761847, 'number': 817} {'precision': 0.5675675675675675, 'recall': 0.5294117647058824, 'f1': 0.5478260869565218, 'number': 119} {'precision': 0.8914233576642335, 'recall': 0.9071494893221913, 'f1': 0.8992176714219972, 'number': 1077} 0.8698 0.8798 0.8748 0.7959

Framework versions

  • Transformers 4.41.1
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
42
Safetensors
Model size
130M 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.

Model tree for chintans/lilt-en-funsd

Finetuned
(46)
this model