update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,129 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-sa-4.0
|
3 |
+
base_model: microsoft/layoutlmv3-base
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- wildreceipt
|
8 |
+
metrics:
|
9 |
+
- precision
|
10 |
+
- recall
|
11 |
+
- f1
|
12 |
+
- accuracy
|
13 |
+
model-index:
|
14 |
+
- name: layoutlmv3-finetuned-wildreceipt
|
15 |
+
results:
|
16 |
+
- task:
|
17 |
+
name: Token Classification
|
18 |
+
type: token-classification
|
19 |
+
dataset:
|
20 |
+
name: wildreceipt
|
21 |
+
type: wildreceipt
|
22 |
+
config: WildReceipt
|
23 |
+
split: test
|
24 |
+
args: WildReceipt
|
25 |
+
metrics:
|
26 |
+
- name: Precision
|
27 |
+
type: precision
|
28 |
+
value: 0.8791872597473915
|
29 |
+
- name: Recall
|
30 |
+
type: recall
|
31 |
+
value: 0.8814865794907089
|
32 |
+
- name: F1
|
33 |
+
type: f1
|
34 |
+
value: 0.8803354182418035
|
35 |
+
- name: Accuracy
|
36 |
+
type: accuracy
|
37 |
+
value: 0.9270261366132221
|
38 |
+
---
|
39 |
+
|
40 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
41 |
+
should probably proofread and complete it, then remove this comment. -->
|
42 |
+
|
43 |
+
# layoutlmv3-finetuned-wildreceipt
|
44 |
+
|
45 |
+
This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the wildreceipt dataset.
|
46 |
+
It achieves the following results on the evaluation set:
|
47 |
+
- Loss: 0.3081
|
48 |
+
- Precision: 0.8792
|
49 |
+
- Recall: 0.8815
|
50 |
+
- F1: 0.8803
|
51 |
+
- Accuracy: 0.9270
|
52 |
+
|
53 |
+
## Model description
|
54 |
+
|
55 |
+
More information needed
|
56 |
+
|
57 |
+
## Intended uses & limitations
|
58 |
+
|
59 |
+
More information needed
|
60 |
+
|
61 |
+
## Training and evaluation data
|
62 |
+
|
63 |
+
More information needed
|
64 |
+
|
65 |
+
## Training procedure
|
66 |
+
|
67 |
+
### Training hyperparameters
|
68 |
+
|
69 |
+
The following hyperparameters were used during training:
|
70 |
+
- learning_rate: 1e-05
|
71 |
+
- train_batch_size: 4
|
72 |
+
- eval_batch_size: 4
|
73 |
+
- seed: 42
|
74 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
75 |
+
- lr_scheduler_type: linear
|
76 |
+
- training_steps: 4000
|
77 |
+
|
78 |
+
### Training results
|
79 |
+
|
80 |
+
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
81 |
+
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
82 |
+
| No log | 0.32 | 100 | 1.3430 | 0.5041 | 0.1959 | 0.2821 | 0.6414 |
|
83 |
+
| No log | 0.63 | 200 | 0.8931 | 0.6739 | 0.5367 | 0.5975 | 0.7786 |
|
84 |
+
| No log | 0.95 | 300 | 0.6793 | 0.7332 | 0.6410 | 0.6840 | 0.8273 |
|
85 |
+
| No log | 1.26 | 400 | 0.5804 | 0.7659 | 0.7090 | 0.7364 | 0.8507 |
|
86 |
+
| 1.0357 | 1.58 | 500 | 0.4876 | 0.7919 | 0.7551 | 0.7731 | 0.8723 |
|
87 |
+
| 1.0357 | 1.89 | 600 | 0.4417 | 0.8009 | 0.7997 | 0.8003 | 0.8857 |
|
88 |
+
| 1.0357 | 2.21 | 700 | 0.3937 | 0.8256 | 0.8200 | 0.8228 | 0.8973 |
|
89 |
+
| 1.0357 | 2.52 | 800 | 0.3904 | 0.8143 | 0.8321 | 0.8231 | 0.8958 |
|
90 |
+
| 1.0357 | 2.84 | 900 | 0.3638 | 0.8462 | 0.8211 | 0.8334 | 0.9010 |
|
91 |
+
| 0.3989 | 3.15 | 1000 | 0.3586 | 0.8386 | 0.8447 | 0.8417 | 0.9055 |
|
92 |
+
| 0.3989 | 3.47 | 1100 | 0.3227 | 0.8382 | 0.8564 | 0.8472 | 0.9104 |
|
93 |
+
| 0.3989 | 3.79 | 1200 | 0.3120 | 0.8538 | 0.8522 | 0.8530 | 0.9119 |
|
94 |
+
| 0.3989 | 4.1 | 1300 | 0.3283 | 0.8498 | 0.8559 | 0.8528 | 0.9117 |
|
95 |
+
| 0.3989 | 4.42 | 1400 | 0.3084 | 0.8595 | 0.8606 | 0.8600 | 0.9165 |
|
96 |
+
| 0.2727 | 4.73 | 1500 | 0.3026 | 0.8552 | 0.8666 | 0.8609 | 0.9159 |
|
97 |
+
| 0.2727 | 5.05 | 1600 | 0.3052 | 0.8633 | 0.8537 | 0.8585 | 0.9165 |
|
98 |
+
| 0.2727 | 5.36 | 1700 | 0.3052 | 0.8505 | 0.8747 | 0.8625 | 0.9165 |
|
99 |
+
| 0.2727 | 5.68 | 1800 | 0.3040 | 0.8579 | 0.8690 | 0.8634 | 0.9164 |
|
100 |
+
| 0.2727 | 5.99 | 1900 | 0.2926 | 0.8717 | 0.8696 | 0.8707 | 0.9205 |
|
101 |
+
| 0.2059 | 6.31 | 2000 | 0.3004 | 0.8646 | 0.8753 | 0.8699 | 0.9207 |
|
102 |
+
| 0.2059 | 6.62 | 2100 | 0.2973 | 0.8711 | 0.8742 | 0.8726 | 0.9215 |
|
103 |
+
| 0.2059 | 6.94 | 2200 | 0.3010 | 0.8650 | 0.8761 | 0.8705 | 0.9214 |
|
104 |
+
| 0.2059 | 7.26 | 2300 | 0.3028 | 0.8654 | 0.8760 | 0.8706 | 0.9214 |
|
105 |
+
| 0.2059 | 7.57 | 2400 | 0.2956 | 0.8769 | 0.8769 | 0.8769 | 0.9260 |
|
106 |
+
| 0.1617 | 7.89 | 2500 | 0.2871 | 0.8746 | 0.8778 | 0.8762 | 0.9266 |
|
107 |
+
| 0.1617 | 8.2 | 2600 | 0.3092 | 0.8632 | 0.8797 | 0.8714 | 0.9226 |
|
108 |
+
| 0.1617 | 8.52 | 2700 | 0.3042 | 0.8834 | 0.8738 | 0.8786 | 0.9265 |
|
109 |
+
| 0.1617 | 8.83 | 2800 | 0.3092 | 0.8672 | 0.8793 | 0.8732 | 0.9224 |
|
110 |
+
| 0.1617 | 9.15 | 2900 | 0.3014 | 0.8738 | 0.8841 | 0.8789 | 0.9256 |
|
111 |
+
| 0.1359 | 9.46 | 3000 | 0.3038 | 0.8763 | 0.8760 | 0.8762 | 0.9249 |
|
112 |
+
| 0.1359 | 9.78 | 3100 | 0.3087 | 0.8730 | 0.8797 | 0.8763 | 0.9241 |
|
113 |
+
| 0.1359 | 10.09 | 3200 | 0.3021 | 0.8740 | 0.8812 | 0.8776 | 0.9251 |
|
114 |
+
| 0.1359 | 10.41 | 3300 | 0.2975 | 0.8790 | 0.8836 | 0.8812 | 0.9268 |
|
115 |
+
| 0.1359 | 10.73 | 3400 | 0.3121 | 0.8734 | 0.8809 | 0.8771 | 0.9254 |
|
116 |
+
| 0.1192 | 11.04 | 3500 | 0.3111 | 0.8812 | 0.8794 | 0.8803 | 0.9260 |
|
117 |
+
| 0.1192 | 11.36 | 3600 | 0.3101 | 0.8785 | 0.8790 | 0.8788 | 0.9261 |
|
118 |
+
| 0.1192 | 11.67 | 3700 | 0.3082 | 0.8790 | 0.8829 | 0.8809 | 0.9275 |
|
119 |
+
| 0.1192 | 11.99 | 3800 | 0.3081 | 0.8822 | 0.8830 | 0.8826 | 0.9276 |
|
120 |
+
| 0.1192 | 12.3 | 3900 | 0.3100 | 0.8800 | 0.8809 | 0.8805 | 0.9269 |
|
121 |
+
| 0.1065 | 12.62 | 4000 | 0.3081 | 0.8792 | 0.8815 | 0.8803 | 0.9270 |
|
122 |
+
|
123 |
+
|
124 |
+
### Framework versions
|
125 |
+
|
126 |
+
- Transformers 4.32.0.dev0
|
127 |
+
- Pytorch 2.0.1+cu118
|
128 |
+
- Datasets 2.14.3
|
129 |
+
- Tokenizers 0.13.3
|