File size: 9,388 Bytes
8821621
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
---
tags:
- generated_from_trainer
datasets:
- funsd
model-index:
- name: layoutlm-funsd
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# layoutlm-funsd

This model is a fine-tuned version of [microsoft/layoutlm-base-uncased](https://huggingface.co/microsoft/layoutlm-base-uncased) on the funsd dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8261
- Answer: {'precision': 0.5727482678983834, 'recall': 0.6131025957972805, 'f1': 0.5922388059701492, 'number': 809}
- Header: {'precision': 0.09302325581395349, 'recall': 0.03361344537815126, 'f1': 0.04938271604938272, 'number': 119}
- Question: {'precision': 0.6384228187919463, 'recall': 0.7145539906103286, 'f1': 0.6743464776251661, 'number': 1065}
- Overall Precision: 0.6002
- Overall Recall: 0.6327
- Overall F1: 0.6160
- Overall Accuracy: 0.7523

## 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: 64
- eval_batch_size: 32
- 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 | Answer                                                                                                         | Header                                                                                                         | Question                                                                                                     | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------:|:-----------------:|:--------------:|:----------:|:----------------:|
| 1.9264        | 1.0   | 3    | 1.7763          | {'precision': 0.011029411764705883, 'recall': 0.022249690976514216, 'f1': 0.01474805407619828, 'number': 809}  | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.09483960948396095, 'recall': 0.12769953051643193, 'f1': 0.108843537414966, 'number': 1065}   | 0.0483            | 0.0773         | 0.0595     | 0.3277           |
| 1.7361        | 2.0   | 6    | 1.6376          | {'precision': 0.0064754856614246065, 'recall': 0.00865265760197775, 'f1': 0.007407407407407408, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.17735470941883769, 'recall': 0.16619718309859155, 'f1': 0.17159476490547748, 'number': 1065} | 0.0885            | 0.0923         | 0.0904     | 0.3852           |
| 1.6212        | 3.0   | 9    | 1.5225          | {'precision': 0.02002002002002002, 'recall': 0.024721878862793572, 'f1': 0.022123893805309734, 'number': 809}  | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.27049180327868855, 'recall': 0.27887323943661974, 'f1': 0.27461858529819694, 'number': 1065} | 0.1512            | 0.1591         | 0.1550     | 0.4422           |
| 1.5178        | 4.0   | 12   | 1.4133          | {'precision': 0.05408388520971302, 'recall': 0.06056860321384425, 'f1': 0.05714285714285715, 'number': 809}    | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.313953488372093, 'recall': 0.38028169014084506, 'f1': 0.34394904458598724, 'number': 1065}   | 0.2067            | 0.2278         | 0.2168     | 0.5062           |
| 1.3853        | 5.0   | 15   | 1.3086          | {'precision': 0.08221024258760108, 'recall': 0.0754017305315204, 'f1': 0.07865892972275951, 'number': 809}     | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.3780202650038971, 'recall': 0.45539906103286387, 'f1': 0.4131175468483816, 'number': 1065}   | 0.2696            | 0.2740         | 0.2718     | 0.5453           |
| 1.2546        | 6.0   | 18   | 1.2110          | {'precision': 0.1463768115942029, 'recall': 0.12484548825710753, 'f1': 0.13475650433622416, 'number': 809}     | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.4551282051282051, 'recall': 0.5333333333333333, 'f1': 0.49113705144833547, 'number': 1065}   | 0.3452            | 0.3357         | 0.3404     | 0.5822           |
| 1.1842        | 7.0   | 21   | 1.1217          | {'precision': 0.2563739376770538, 'recall': 0.22373300370828184, 'f1': 0.23894389438943894, 'number': 809}     | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.4908512330946698, 'recall': 0.5793427230046948, 'f1': 0.5314384151593453, 'number': 1065}    | 0.4055            | 0.4004         | 0.4029     | 0.6223           |
| 1.0564        | 8.0   | 24   | 1.0490          | {'precision': 0.364461738002594, 'recall': 0.3473423980222497, 'f1': 0.3556962025316456, 'number': 809}        | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.509976057462091, 'recall': 0.6, 'f1': 0.551337359792925, 'number': 1065}                     | 0.4523            | 0.4616         | 0.4569     | 0.6679           |
| 0.9865        | 9.0   | 27   | 0.9863          | {'precision': 0.4305555555555556, 'recall': 0.4215080346106304, 'f1': 0.42598376014990635, 'number': 809}      | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}                                                    | {'precision': 0.5528726061615321, 'recall': 0.6234741784037559, 'f1': 0.5860547219770521, 'number': 1065}    | 0.4978            | 0.5043         | 0.5010     | 0.6986           |
| 0.9281        | 10.0  | 30   | 0.9357          | {'precision': 0.49454545454545457, 'recall': 0.5043263288009888, 'f1': 0.49938800489596086, 'number': 809}     | {'precision': 0.034482758620689655, 'recall': 0.008403361344537815, 'f1': 0.013513513513513513, 'number': 119} | {'precision': 0.5873287671232876, 'recall': 0.644131455399061, 'f1': 0.6144200626959248, 'number': 1065}     | 0.5415            | 0.5494         | 0.5455     | 0.7197           |
| 0.8646        | 11.0  | 33   | 0.8968          | {'precision': 0.5333333333333333, 'recall': 0.5438813349814586, 'f1': 0.5385556915544676, 'number': 809}       | {'precision': 0.0625, 'recall': 0.01680672268907563, 'f1': 0.026490066225165563, 'number': 119}                | {'precision': 0.6031746031746031, 'recall': 0.6779342723004694, 'f1': 0.6383731211317418, 'number': 1065}    | 0.5667            | 0.5840         | 0.5752     | 0.7344           |
| 0.828         | 12.0  | 36   | 0.8653          | {'precision': 0.5617577197149644, 'recall': 0.584672435105068, 'f1': 0.5729860690490611, 'number': 809}        | {'precision': 0.07692307692307693, 'recall': 0.025210084033613446, 'f1': 0.0379746835443038, 'number': 119}    | {'precision': 0.6204013377926422, 'recall': 0.6967136150234742, 'f1': 0.6563467492260062, 'number': 1065}    | 0.5864            | 0.6111         | 0.5985     | 0.7442           |
| 0.7803        | 13.0  | 39   | 0.8442          | {'precision': 0.5667828106852497, 'recall': 0.6032138442521632, 'f1': 0.5844311377245508, 'number': 809}       | {'precision': 0.07142857142857142, 'recall': 0.025210084033613446, 'f1': 0.037267080745341616, 'number': 119}  | {'precision': 0.6343906510851419, 'recall': 0.7136150234741784, 'f1': 0.6716747680070703, 'number': 1065}    | 0.5954            | 0.6277         | 0.6111     | 0.7504           |
| 0.771         | 14.0  | 42   | 0.8312          | {'precision': 0.5679723502304147, 'recall': 0.6093943139678616, 'f1': 0.5879546809779368, 'number': 809}       | {'precision': 0.09302325581395349, 'recall': 0.03361344537815126, 'f1': 0.04938271604938272, 'number': 119}    | {'precision': 0.6376569037656904, 'recall': 0.7154929577464789, 'f1': 0.6743362831858407, 'number': 1065}    | 0.5978            | 0.6317         | 0.6143     | 0.7516           |
| 0.7843        | 15.0  | 45   | 0.8261          | {'precision': 0.5727482678983834, 'recall': 0.6131025957972805, 'f1': 0.5922388059701492, 'number': 809}       | {'precision': 0.09302325581395349, 'recall': 0.03361344537815126, 'f1': 0.04938271604938272, 'number': 119}    | {'precision': 0.6384228187919463, 'recall': 0.7145539906103286, 'f1': 0.6743464776251661, 'number': 1065}    | 0.6002            | 0.6327         | 0.6160     | 0.7523           |


### Framework versions

- Transformers 4.22.0.dev0
- Pytorch 1.12.1+cu116
- Datasets 2.4.0
- Tokenizers 0.12.1