Hugging Face's logo

Luciano
/
xlm-roberta-base-finetuned-lener-br

Token Classification
Transformers
PyTorch
TensorBoard
Portuguese
xlm-roberta
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
3
Luciano commited on
Commit
bb23b4f
•
1 Parent(s): 8625c9c

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -110
README.md CHANGED
@@ -24,103 +24,16 @@ model-index:
24
  metrics:
25
  - name: Precision
26
  type: precision
27
- value: 0.7563938618925832
28
  - name: Recall
29
  type: recall
30
- value: 0.9172912897389507
31
  - name: F1
32
  type: f1
33
- value: 0.8291087489779232
34
  - name: Accuracy
35
  type: accuracy
36
- value: 0.9672628386152076
37
- - task:
38
- type: token-classification
39
- name: Token Classification
40
- dataset:
41
- name: lener_br
42
- type: lener_br
43
- config: lener_br
44
- split: test
45
- metrics:
46
- - name: Accuracy
47
- type: accuracy
48
- value: 0.9856250947530907
49
- verified: true
50
- - name: Precision
51
- type: precision
52
- value: 0.9880955923930835
53
- verified: true
54
- - name: Recall
55
- type: recall
56
- value: 0.988198526676092
57
- verified: true
58
- - name: F1
59
- type: f1
60
- value: 0.9881470568539475
61
- verified: true
62
- - name: loss
63
- type: loss
64
- value: 0.110235795378685
65
- verified: true
66
- - task:
67
- type: token-classification
68
- name: Token Classification
69
- dataset:
70
- name: lener_br
71
- type: lener_br
72
- config: lener_br
73
- split: validation
74
- metrics:
75
- - name: Accuracy
76
- type: accuracy
77
- value: 0.9672777802680532
78
- verified: true
79
- - name: Precision
80
- type: precision
81
- value: 0.9857782940590774
82
- verified: true
83
- - name: Recall
84
- type: recall
85
- value: 0.9705896769766188
86
- verified: true
87
- - name: F1
88
- type: f1
89
- value: 0.9781250257279995
90
- verified: true
91
- - name: loss
92
- type: loss
93
- value: 0.24753354489803314
94
- verified: true
95
- - task:
96
- type: token-classification
97
- name: Token Classification
98
- dataset:
99
- name: lener_br
100
- type: lener_br
101
- config: lener_br
102
- split: train
103
- metrics:
104
- - name: Accuracy
105
- type: accuracy
106
- value: 0.999662705763308
107
- verified: true
108
- - name: Precision
109
- type: precision
110
- value: 0.9996394685766821
111
- verified: true
112
- - name: Recall
113
- type: recall
114
- value: 0.9997265731807351
115
- verified: true
116
- - name: F1
117
- type: f1
118
- value: 0.9996830189813041
119
- verified: true
120
- - name: loss
121
- type: loss
122
- value: 0.0008438663207925856
123
- verified: true
124
  ---
125
 
126
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -131,10 +44,10 @@ should probably proofread and complete it, then remove this comment. -->
131
  This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the lener_br dataset.
132
  It achieves the following results on the evaluation set:
133
  - Loss: nan
134
- - Precision: 0.7564
135
- - Recall: 0.9173
136
- - F1: 0.8291
137
- - Accuracy: 0.9673
138
 
139
  ## Model description
140
 
@@ -160,26 +73,27 @@ The following hyperparameters were used during training:
160
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
161
  - lr_scheduler_type: linear
162
  - num_epochs: 15
 
163
 
164
  ### Training results
165
 
166
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
167
  |:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
168
- | 0.0939 | 1.0 | 1957 | nan | 0.6718 | 0.8545 | 0.7522 | 0.9609 |
169
- | 0.0462 | 2.0 | 3914 | nan | 0.7637 | 0.8739 | 0.8151 | 0.9657 |
170
- | 0.0286 | 3.0 | 5871 | nan | 0.7357 | 0.9077 | 0.8127 | 0.9691 |
171
- | 0.0253 | 4.0 | 7828 | nan | 0.7497 | 0.8989 | 0.8176 | 0.9690 |
172
- | 0.0209 | 5.0 | 9785 | nan | 0.7363 | 0.9196 | 0.8178 | 0.9624 |
173
- | 0.0149 | 6.0 | 11742 | nan | 0.7209 | 0.9201 | 0.8084 | 0.9673 |
174
- | 0.0149 | 7.0 | 13699 | nan | 0.7508 | 0.8987 | 0.8181 | 0.9682 |
175
- | 0.0099 | 8.0 | 15656 | nan | 0.7837 | 0.8692 | 0.8243 | 0.9617 |
176
- | 0.0067 | 9.0 | 17613 | nan | 0.8086 | 0.8638 | 0.8353 | 0.9703 |
177
- | 0.0046 | 10.0 | 19570 | nan | 0.7518 | 0.9209 | 0.8278 | 0.9682 |
178
- | 0.0047 | 11.0 | 21527 | nan | 0.7504 | 0.9101 | 0.8226 | 0.9681 |
179
- | 0.002 | 12.0 | 23484 | nan | 0.7890 | 0.9082 | 0.8444 | 0.9646 |
180
- | 0.0033 | 13.0 | 25441 | nan | 0.7629 | 0.9157 | 0.8324 | 0.9675 |
181
- | 0.0029 | 14.0 | 27398 | nan | 0.7484 | 0.9155 | 0.8235 | 0.9658 |
182
- | 0.0009 | 15.0 | 29355 | nan | 0.7564 | 0.9173 | 0.8291 | 0.9673 |
183
 
184
 
185
  ### Framework versions
 
24
  metrics:
25
  - name: Precision
26
  type: precision
27
+ value: 0.844312854675549
28
  - name: Recall
29
  type: recall
30
+ value: 0.8844662703540966
31
  - name: F1
32
  type: f1
33
+ value: 0.8639232517041151
34
  - name: Accuracy
35
  type: accuracy
36
+ value: 0.97516697297055
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
37
  ---
38
 
39
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
44
  This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the lener_br dataset.
45
  It achieves the following results on the evaluation set:
46
  - Loss: nan
47
+ - Precision: 0.8443
48
+ - Recall: 0.8845
49
+ - F1: 0.8639
50
+ - Accuracy: 0.9752
51
 
52
  ## Model description
53
 
 
73
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
74
  - lr_scheduler_type: linear
75
  - num_epochs: 15
76
+ - mixed_precision_training: Native AMP
77
 
78
  ### Training results
79
 
80
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
81
  |:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
82
+ | 0.0832 | 1.0 | 1957 | nan | 0.6752 | 0.8625 | 0.7575 | 0.9578 |
83
+ | 0.0477 | 2.0 | 3914 | nan | 0.8391 | 0.8839 | 0.8609 | 0.9704 |
84
+ | 0.029 | 3.0 | 5871 | nan | 0.7530 | 0.9059 | 0.8224 | 0.9648 |
85
+ | 0.0223 | 4.0 | 7828 | nan | 0.7488 | 0.8744 | 0.8067 | 0.9659 |
86
+ | 0.0234 | 5.0 | 9785 | nan | 0.7216 | 0.8783 | 0.7923 | 0.9644 |
87
+ | 0.0171 | 6.0 | 11742 | nan | 0.7072 | 0.8969 | 0.7908 | 0.9642 |
88
+ | 0.0121 | 7.0 | 13699 | nan | 0.7769 | 0.8775 | 0.8241 | 0.9681 |
89
+ | 0.0093 | 8.0 | 15656 | nan | 0.7218 | 0.8772 | 0.7920 | 0.9621 |
90
+ | 0.0074 | 9.0 | 17613 | nan | 0.8241 | 0.8767 | 0.8496 | 0.9739 |
91
+ | 0.0055 | 10.0 | 19570 | nan | 0.7369 | 0.8801 | 0.8021 | 0.9638 |
92
+ | 0.0055 | 11.0 | 21527 | nan | 0.8443 | 0.8845 | 0.8639 | 0.9752 |
93
+ | 0.0029 | 12.0 | 23484 | nan | 0.8338 | 0.8935 | 0.8626 | 0.9753 |
94
+ | 0.0026 | 13.0 | 25441 | nan | 0.7721 | 0.8992 | 0.8308 | 0.9694 |
95
+ | 0.004 | 14.0 | 27398 | nan | 0.7466 | 0.8886 | 0.8114 | 0.9672 |
96
+ | 0.0006 | 15.0 | 29355 | nan | 0.7518 | 0.8995 | 0.8190 | 0.9686 |
97
 
98
 
99
  ### Framework versions