Hugging Face's logo

HueyNemud
/
icdar23-entrydetector_plaintext_breaks_indents_left_diff

Token Classification
Transformers
PyTorch
camembert
Generated from Trainer
Inference Endpoints
Model card Files Community
HueyNemud commited on
Commit
4d10314
•
1 Parent(s): 8a79633

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +34 -18
README.md CHANGED
@@ -13,13 +13,13 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  This model is a fine-tuned version of [HueyNemud/das22-10-camembert_pretrained](https://huggingface.co/HueyNemud/das22-10-camembert_pretrained) on the None dataset.
15
  It achieves the following results on the evaluation set:
16
- - Loss: 0.0091
17
- - Ebegin: {'precision': 0.9909729187562688, 'recall': 0.9873417721518988, 'f1': 0.9891540130151844, 'number': 3002}
18
- - Eend: {'precision': 0.986648865153538, 'recall': 0.9853333333333333, 'f1': 0.9859906604402935, 'number': 3000}
19
- - Overall Precision: 0.9888
20
- - Overall Recall: 0.9863
21
- - Overall F1: 0.9876
22
- - Overall Accuracy: 0.9979
23
 
24
  ## Model description
25
 
@@ -44,26 +44,42 @@ The following hyperparameters were used during training:
44
  - seed: 42
45
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
  - lr_scheduler_type: linear
47
- - training_steps: 6000
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
52
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
53
- | No log | 0.07 | 300 | 0.0315 | 0.9611 | 0.9872 | 0.9740 | 0.9956 |
54
- | 0.1635 | 0.14 | 600 | 0.0130 | 0.9850 | 0.9908 | 0.9879 | 0.9979 |
55
- | 0.1635 | 0.21 | 900 | 0.0096 | 0.9818 | 0.9951 | 0.9884 | 0.9979 |
56
- | 0.0194 | 0.29 | 1200 | 0.0074 | 0.9888 | 0.9908 | 0.9898 | 0.9982 |
57
- | 0.0107 | 0.36 | 1500 | 0.0062 | 0.9885 | 0.9943 | 0.9914 | 0.9984 |
58
- | 0.0107 | 0.43 | 1800 | 0.0082 | 0.9928 | 0.9870 | 0.9899 | 0.9982 |
59
- | 0.0078 | 0.5 | 2100 | 0.0060 | 0.9860 | 0.9948 | 0.9904 | 0.9983 |
60
- | 0.0078 | 0.57 | 2400 | 0.0064 | 0.9865 | 0.9941 | 0.9903 | 0.9983 |
61
- | 0.0061 | 0.64 | 2700 | 0.0055 | 0.9938 | 0.9876 | 0.9907 | 0.9983 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
62
 
63
 
64
  ### Framework versions
65
 
66
- - Transformers 4.26.0
67
  - Pytorch 1.13.1+cu116
68
  - Datasets 2.9.0
69
  - Tokenizers 0.13.2
 
13
 
14
  This model is a fine-tuned version of [HueyNemud/das22-10-camembert_pretrained](https://huggingface.co/HueyNemud/das22-10-camembert_pretrained) on the None dataset.
15
  It achieves the following results on the evaluation set:
16
+ - Loss: 0.0104
17
+ - Ebegin: {'precision': 0.9969477298740939, 'recall': 0.9827002632568634, 'f1': 0.9897727272727271, 'number': 2659}
18
+ - Eend: {'precision': 0.9988553987027852, 'recall': 0.9783258594917787, 'f1': 0.9884840475740986, 'number': 2676}
19
+ - Overall Precision: 0.9979
20
+ - Overall Recall: 0.9805
21
+ - Overall F1: 0.9891
22
+ - Overall Accuracy: 0.9981
23
 
24
  ## Model description
25
 
 
44
  - seed: 42
45
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
  - lr_scheduler_type: linear
47
+ - training_steps: 7500
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
52
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
53
+ | No log | 0.07 | 300 | 0.0368 | 0.9643 | 0.9797 | 0.9719 | 0.9960 |
54
+ | 0.1994 | 0.14 | 600 | 0.0176 | 0.9847 | 0.9854 | 0.9851 | 0.9976 |
55
+ | 0.1994 | 0.21 | 900 | 0.0137 | 0.9926 | 0.9800 | 0.9862 | 0.9976 |
56
+ | 0.0225 | 0.29 | 1200 | 0.0102 | 0.9900 | 0.9901 | 0.9901 | 0.9983 |
57
+ | 0.0124 | 0.36 | 1500 | 0.0075 | 0.9930 | 0.9909 | 0.9919 | 0.9985 |
58
+ | 0.0124 | 0.43 | 1800 | 0.0065 | 0.9947 | 0.9905 | 0.9926 | 0.9987 |
59
+ | 0.0093 | 0.5 | 2100 | 0.0066 | 0.9952 | 0.9903 | 0.9928 | 0.9987 |
60
+ | 0.0093 | 0.57 | 2400 | 0.0068 | 0.9897 | 0.9935 | 0.9916 | 0.9985 |
61
+ | 0.0095 | 0.64 | 2700 | 0.0064 | 0.9938 | 0.9907 | 0.9923 | 0.9986 |
62
+ | 0.0068 | 0.72 | 3000 | 0.0051 | 0.9927 | 0.9935 | 0.9931 | 0.9988 |
63
+ | 0.0068 | 0.79 | 3300 | 0.0050 | 0.9907 | 0.9948 | 0.9927 | 0.9987 |
64
+ | 0.0064 | 0.86 | 3600 | 0.0064 | 0.9881 | 0.9955 | 0.9918 | 0.9985 |
65
+ | 0.0064 | 0.93 | 3900 | 0.0073 | 0.9895 | 0.9910 | 0.9902 | 0.9983 |
66
+ | 0.0057 | 1.0 | 4200 | 0.0051 | 0.9966 | 0.9896 | 0.9931 | 0.9988 |
67
+ | 0.0042 | 1.07 | 4500 | 0.0050 | 0.9943 | 0.9935 | 0.9939 | 0.9989 |
68
+ | 0.0042 | 1.14 | 4800 | 0.0050 | 0.9966 | 0.9903 | 0.9934 | 0.9988 |
69
+ | 0.0039 | 1.22 | 5100 | 0.0050 | 0.9947 | 0.9934 | 0.9940 | 0.9989 |
70
+ | 0.0039 | 1.29 | 5400 | 0.0049 | 0.9941 | 0.9931 | 0.9936 | 0.9989 |
71
+ | 0.0034 | 1.36 | 5700 | 0.0043 | 0.9966 | 0.9914 | 0.9940 | 0.9989 |
72
+ | 0.0034 | 1.43 | 6000 | 0.0045 | 0.9958 | 0.9920 | 0.9939 | 0.9989 |
73
+ | 0.0034 | 1.5 | 6300 | 0.0043 | 0.9935 | 0.9935 | 0.9935 | 0.9988 |
74
+ | 0.002 | 1.57 | 6600 | 0.0044 | 0.9953 | 0.9934 | 0.9944 | 0.9990 |
75
+ | 0.002 | 1.65 | 6900 | 0.0044 | 0.9952 | 0.9929 | 0.9940 | 0.9989 |
76
+ | 0.0037 | 1.72 | 7200 | 0.0041 | 0.9952 | 0.9931 | 0.9941 | 0.9990 |
77
+ | 0.0028 | 1.79 | 7500 | 0.0042 | 0.9958 | 0.9930 | 0.9944 | 0.9990 |
78
 
79
 
80
  ### Framework versions
81
 
82
+ - Transformers 4.26.1
83
  - Pytorch 1.13.1+cu116
84
  - Datasets 2.9.0
85
  - Tokenizers 0.13.2