Hugging Face's logo

henryscheible
/
bert-large-uncased_crows_pairs_finetuned

Text Classification
Transformers
PyTorch
bert
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
2
henryscheible commited on
Commit
b4e38fe
•
1 Parent(s): 3a8b1f1

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +63 -26
README.md CHANGED
@@ -21,7 +21,7 @@ model-index:
21
  metrics:
22
  - name: Accuracy
23
  type: accuracy
24
- value: 0.4900662251655629
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -31,8 +31,8 @@ should probably proofread and complete it, then remove this comment. -->
31
 
32
  This model is a fine-tuned version of [bert-large-uncased](https://huggingface.co/bert-large-uncased) on the crows_pairs dataset.
33
  It achieves the following results on the evaluation set:
34
- - Loss: 0.6934
35
- - Accuracy: 0.4901
36
 
37
  ## Model description
38
 
@@ -51,38 +51,75 @@ More information needed
51
  ### Training hyperparameters
52
 
53
  The following hyperparameters were used during training:
54
- - learning_rate: 0.01
55
- - train_batch_size: 128
56
  - eval_batch_size: 64
57
  - seed: 42
58
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
  - lr_scheduler_type: linear
60
- - num_epochs: 20
61
 
62
  ### Training results
63
 
64
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
65
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
66
- | No log | 1.0 | 10 | 1.6646 | 0.4901 |
67
- | No log | 2.0 | 20 | 0.7628 | 0.5099 |
68
- | No log | 3.0 | 30 | 2.1840 | 0.5099 |
69
- | No log | 4.0 | 40 | 0.8822 | 0.4901 |
70
- | No log | 5.0 | 50 | 0.9210 | 0.4901 |
71
- | No log | 6.0 | 60 | 0.7963 | 0.4901 |
72
- | No log | 7.0 | 70 | 0.8562 | 0.5099 |
73
- | No log | 8.0 | 80 | 0.8075 | 0.5099 |
74
- | No log | 9.0 | 90 | 1.0372 | 0.4901 |
75
- | No log | 10.0 | 100 | 0.7340 | 0.5099 |
76
- | No log | 11.0 | 110 | 0.7658 | 0.5099 |
77
- | No log | 12.0 | 120 | 0.8453 | 0.4901 |
78
- | No log | 13.0 | 130 | 0.7098 | 0.5099 |
79
- | No log | 14.0 | 140 | 0.7295 | 0.5099 |
80
- | No log | 15.0 | 150 | 0.7492 | 0.5099 |
81
- | No log | 16.0 | 160 | 0.7811 | 0.4901 |
82
- | No log | 17.0 | 170 | 0.6976 | 0.4901 |
83
- | No log | 18.0 | 180 | 0.6936 | 0.5099 |
84
- | No log | 19.0 | 190 | 0.6934 | 0.4901 |
85
- | No log | 20.0 | 200 | 0.6934 | 0.4901 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
86
 
87
 
88
  ### Framework versions
 
21
  metrics:
22
  - name: Accuracy
23
  type: accuracy
24
+ value: 0.46688741721854304
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
31
 
32
  This model is a fine-tuned version of [bert-large-uncased](https://huggingface.co/bert-large-uncased) on the crows_pairs dataset.
33
  It achieves the following results on the evaluation set:
34
+ - Loss: 0.6936
35
+ - Accuracy: 0.4669
36
 
37
  ## Model description
38
 
 
51
  ### Training hyperparameters
52
 
53
  The following hyperparameters were used during training:
54
+ - learning_rate: 0.0005
55
+ - train_batch_size: 64
56
  - eval_batch_size: 64
57
  - seed: 42
58
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
  - lr_scheduler_type: linear
60
+ - num_epochs: 30
61
 
62
  ### Training results
63
 
64
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
65
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
66
+ | 0.8146 | 0.53 | 10 | 0.6914 | 0.5331 |
67
+ | 0.7104 | 1.05 | 20 | 0.6910 | 0.5331 |
68
+ | 0.7078 | 1.58 | 30 | 0.7292 | 0.4669 |
69
+ | 0.716 | 2.11 | 40 | 0.7033 | 0.4669 |
70
+ | 0.7273 | 2.63 | 50 | 0.6946 | 0.5331 |
71
+ | 0.7285 | 3.16 | 60 | 0.6983 | 0.5331 |
72
+ | 0.7244 | 3.68 | 70 | 0.6958 | 0.5331 |
73
+ | 0.7283 | 4.21 | 80 | 0.7013 | 0.4669 |
74
+ | 0.7131 | 4.74 | 90 | 0.7063 | 0.4669 |
75
+ | 0.7144 | 5.26 | 100 | 0.7149 | 0.4669 |
76
+ | 0.7237 | 5.79 | 110 | 0.6913 | 0.5331 |
77
+ | 0.7074 | 6.32 | 120 | 0.6922 | 0.5331 |
78
+ | 0.7034 | 6.84 | 130 | 0.6910 | 0.5331 |
79
+ | 0.699 | 7.37 | 140 | 0.7251 | 0.4669 |
80
+ | 0.7183 | 7.89 | 150 | 0.7216 | 0.5331 |
81
+ | 0.7106 | 8.42 | 160 | 0.7046 | 0.4669 |
82
+ | 0.7107 | 8.95 | 170 | 0.6923 | 0.5331 |
83
+ | 0.6963 | 9.47 | 180 | 0.7056 | 0.4669 |
84
+ | 0.7068 | 10.0 | 190 | 0.6911 | 0.5331 |
85
+ | 0.7088 | 10.53 | 200 | 0.6963 | 0.4669 |
86
+ | 0.7074 | 11.05 | 210 | 0.7269 | 0.4669 |
87
+ | 0.7233 | 11.58 | 220 | 0.6995 | 0.5331 |
88
+ | 0.7261 | 12.11 | 230 | 0.6921 | 0.5331 |
89
+ | 0.6997 | 12.63 | 240 | 0.6971 | 0.4669 |
90
+ | 0.6993 | 13.16 | 250 | 0.7103 | 0.4669 |
91
+ | 0.7073 | 13.68 | 260 | 0.6923 | 0.5331 |
92
+ | 0.697 | 14.21 | 270 | 0.6938 | 0.4669 |
93
+ | 0.7057 | 14.74 | 280 | 0.6948 | 0.5331 |
94
+ | 0.7165 | 15.26 | 290 | 0.7053 | 0.4669 |
95
+ | 0.7172 | 15.79 | 300 | 0.6910 | 0.5331 |
96
+ | 0.7152 | 16.32 | 310 | 0.6921 | 0.5331 |
97
+ | 0.7115 | 16.84 | 320 | 0.7050 | 0.4669 |
98
+ | 0.7202 | 17.37 | 330 | 0.6911 | 0.5331 |
99
+ | 0.7069 | 17.89 | 340 | 0.6952 | 0.4669 |
100
+ | 0.7061 | 18.42 | 350 | 0.6914 | 0.5331 |
101
+ | 0.7023 | 18.95 | 360 | 0.6943 | 0.4669 |
102
+ | 0.7045 | 19.47 | 370 | 0.6911 | 0.5331 |
103
+ | 0.7114 | 20.0 | 380 | 0.6925 | 0.5331 |
104
+ | 0.6922 | 20.53 | 390 | 0.6910 | 0.5331 |
105
+ | 0.7097 | 21.05 | 400 | 0.6919 | 0.5331 |
106
+ | 0.7142 | 21.58 | 410 | 0.6946 | 0.4669 |
107
+ | 0.7113 | 22.11 | 420 | 0.6933 | 0.4669 |
108
+ | 0.6979 | 22.63 | 430 | 0.6934 | 0.5331 |
109
+ | 0.7214 | 23.16 | 440 | 0.7112 | 0.4669 |
110
+ | 0.6974 | 23.68 | 450 | 0.6929 | 0.5331 |
111
+ | 0.7077 | 24.21 | 460 | 0.6918 | 0.5331 |
112
+ | 0.7123 | 24.74 | 470 | 0.7006 | 0.4669 |
113
+ | 0.7065 | 25.26 | 480 | 0.6978 | 0.4669 |
114
+ | 0.7079 | 25.79 | 490 | 0.6922 | 0.5331 |
115
+ | 0.7063 | 26.32 | 500 | 0.6991 | 0.4669 |
116
+ | 0.7182 | 26.84 | 510 | 0.6956 | 0.4669 |
117
+ | 0.7061 | 27.37 | 520 | 0.6914 | 0.5331 |
118
+ | 0.7069 | 27.89 | 530 | 0.6912 | 0.5331 |
119
+ | 0.7024 | 28.42 | 540 | 0.6929 | 0.5331 |
120
+ | 0.701 | 28.95 | 550 | 0.6954 | 0.4669 |
121
+ | 0.7131 | 29.47 | 560 | 0.6942 | 0.4669 |
122
+ | 0.6999 | 30.0 | 570 | 0.6936 | 0.4669 |
123
 
124
 
125
  ### Framework versions