layoutlm-funsd / README.md
Neha-CanWill's picture
End of training
d21a17a verified
|
raw
history blame
9.28 kB
metadata
license: mit
base_model: microsoft/layoutlm-base-uncased
tags:
  - generated_from_trainer
datasets:
  - funsd
model-index:
  - name: layoutlm-funsd
    results: []

layoutlm-funsd

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

  • Loss: 1.0436
  • Answer: {'precision': 0.3978685612788632, 'recall': 0.553770086526576, 'f1': 0.4630490956072351, 'number': 809}
  • Header: {'precision': 0.32926829268292684, 'recall': 0.226890756302521, 'f1': 0.26865671641791045, 'number': 119}
  • Question: {'precision': 0.5241157556270096, 'recall': 0.612206572769953, 'f1': 0.5647466435686443, 'number': 1065}
  • Overall Precision: 0.4596
  • Overall Recall: 0.5655
  • Overall F1: 0.5071
  • Overall Accuracy: 0.6267

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: 16
  • 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

Training results

Training Loss Epoch Step Validation Loss Answer Header Question Overall Precision Overall Recall Overall F1 Overall Accuracy
1.7148 1.0 10 1.5016 {'precision': 0.08819018404907976, 'recall': 0.14215080346106304, 'f1': 0.10884997633696165, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.2198581560283688, 'recall': 0.08732394366197183, 'f1': 0.125, 'number': 1065} 0.1204 0.1044 0.1118 0.3613
1.4202 2.0 20 1.3572 {'precision': 0.21160042964554243, 'recall': 0.48702101359703337, 'f1': 0.29502059153874954, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.24895977808599168, 'recall': 0.3370892018779343, 'f1': 0.28639808536098926, 'number': 1065} 0.2265 0.3778 0.2832 0.4216
1.2863 3.0 30 1.2150 {'precision': 0.25656167979002625, 'recall': 0.48331273176761436, 'f1': 0.33519074153450495, 'number': 809} {'precision': 0.06779661016949153, 'recall': 0.03361344537815126, 'f1': 0.0449438202247191, 'number': 119} {'precision': 0.3437908496732026, 'recall': 0.49389671361502346, 'f1': 0.4053949903660886, 'number': 1065} 0.2959 0.4621 0.3608 0.4790
1.1633 4.0 40 1.1144 {'precision': 0.2625454545454545, 'recall': 0.446229913473424, 'f1': 0.3305860805860806, 'number': 809} {'precision': 0.3253012048192771, 'recall': 0.226890756302521, 'f1': 0.26732673267326734, 'number': 119} {'precision': 0.37986577181208053, 'recall': 0.5314553990610329, 'f1': 0.4430528375733855, 'number': 1065} 0.3236 0.4787 0.3862 0.5442
1.0585 5.0 50 1.0827 {'precision': 0.3039940828402367, 'recall': 0.5080346106304079, 'f1': 0.38037945395650163, 'number': 809} {'precision': 0.32432432432432434, 'recall': 0.20168067226890757, 'f1': 0.24870466321243526, 'number': 119} {'precision': 0.4149933065595716, 'recall': 0.5821596244131455, 'f1': 0.48456428292301673, 'number': 1065} 0.3613 0.5294 0.4295 0.5700
0.9987 6.0 60 1.0373 {'precision': 0.326783114992722, 'recall': 0.5550061804697157, 'f1': 0.4113605130554283, 'number': 809} {'precision': 0.4074074074074074, 'recall': 0.18487394957983194, 'f1': 0.2543352601156069, 'number': 119} {'precision': 0.453125, 'recall': 0.5173708920187794, 'f1': 0.4831214379658045, 'number': 1065} 0.3865 0.5128 0.4408 0.6016
0.9315 7.0 70 1.0055 {'precision': 0.34718100890207715, 'recall': 0.4338689740420272, 'f1': 0.3857142857142857, 'number': 809} {'precision': 0.3229166666666667, 'recall': 0.2605042016806723, 'f1': 0.28837209302325584, 'number': 119} {'precision': 0.4558011049723757, 'recall': 0.6197183098591549, 'f1': 0.5252686032630322, 'number': 1065} 0.4078 0.5228 0.4582 0.6164
0.8716 8.0 80 1.0112 {'precision': 0.33733013589128696, 'recall': 0.5216316440049443, 'f1': 0.40970873786407763, 'number': 809} {'precision': 0.3717948717948718, 'recall': 0.24369747899159663, 'f1': 0.29441624365482233, 'number': 119} {'precision': 0.44542372881355935, 'recall': 0.6169014084507042, 'f1': 0.5173228346456693, 'number': 1065} 0.3951 0.5559 0.4620 0.6153
0.8102 9.0 90 1.0152 {'precision': 0.3773062730627306, 'recall': 0.5055624227441285, 'f1': 0.4321183306920232, 'number': 809} {'precision': 0.3611111111111111, 'recall': 0.2184873949579832, 'f1': 0.27225130890052357, 'number': 119} {'precision': 0.4880860876249039, 'recall': 0.596244131455399, 'f1': 0.536770921386306, 'number': 1065} 0.4355 0.5369 0.4809 0.6226
0.8003 10.0 100 1.0342 {'precision': 0.3804878048780488, 'recall': 0.5784919653893696, 'f1': 0.45904855321235905, 'number': 809} {'precision': 0.32, 'recall': 0.20168067226890757, 'f1': 0.24742268041237112, 'number': 119} {'precision': 0.5183887915936952, 'recall': 0.5558685446009389, 'f1': 0.5364748527412777, 'number': 1065} 0.4430 0.5439 0.4883 0.6143
0.728 11.0 110 1.0330 {'precision': 0.3871559633027523, 'recall': 0.5216316440049443, 'f1': 0.4444444444444445, 'number': 809} {'precision': 0.29213483146067415, 'recall': 0.2184873949579832, 'f1': 0.25, 'number': 119} {'precision': 0.4981791697013838, 'recall': 0.6422535211267606, 'f1': 0.561115668580804, 'number': 1065} 0.4436 0.5680 0.4981 0.6221
0.7175 12.0 120 1.0841 {'precision': 0.38127090301003347, 'recall': 0.5636588380716935, 'f1': 0.45486284289276807, 'number': 809} {'precision': 0.3684210526315789, 'recall': 0.23529411764705882, 'f1': 0.28717948717948716, 'number': 119} {'precision': 0.5153225806451613, 'recall': 0.6, 'f1': 0.5544468546637744, 'number': 1065} 0.4471 0.5635 0.4986 0.6243
0.6893 13.0 130 1.0501 {'precision': 0.3815126050420168, 'recall': 0.5611866501854141, 'f1': 0.4542271135567784, 'number': 809} {'precision': 0.30952380952380953, 'recall': 0.2184873949579832, 'f1': 0.2561576354679803, 'number': 119} {'precision': 0.5256950294860994, 'recall': 0.5859154929577465, 'f1': 0.5541740674955595, 'number': 1065} 0.4486 0.5539 0.4957 0.6228
0.653 14.0 140 1.0222 {'precision': 0.39345794392523364, 'recall': 0.5203955500618047, 'f1': 0.4481106971793507, 'number': 809} {'precision': 0.34615384615384615, 'recall': 0.226890756302521, 'f1': 0.27411167512690354, 'number': 119} {'precision': 0.5045180722891566, 'recall': 0.6291079812206573, 'f1': 0.55996656916005, 'number': 1065} 0.4515 0.5610 0.5003 0.6269
0.6494 15.0 150 1.0436 {'precision': 0.3978685612788632, 'recall': 0.553770086526576, 'f1': 0.4630490956072351, 'number': 809} {'precision': 0.32926829268292684, 'recall': 0.226890756302521, 'f1': 0.26865671641791045, 'number': 119} {'precision': 0.5241157556270096, 'recall': 0.612206572769953, 'f1': 0.5647466435686443, 'number': 1065} 0.4596 0.5655 0.5071 0.6267

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2