Hugging Face's logo

madatnlp
/
kor-math-roberta-finetune

Fill-Mask
Transformers
TensorFlow
roberta
generated_from_keras_callback
Inference Endpoints
Model card Files Community
madatnlp commited on
Commit
b9f98a2
·
1 Parent(s): b684420

Training in progress epoch 0

Browse files
Files changed (2) hide show
  1. README.md +5 -50
  2. tf_model.h5 +1 -1
README.md CHANGED
@@ -13,9 +13,9 @@ probably proofread and complete it, then remove this comment. -->
13
 
14
  This model is a fine-tuned version of [klue/roberta-base](https://huggingface.co/klue/roberta-base) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
- - Train Loss: 0.7393
17
- - Validation Loss: 0.7801
18
- - Epoch: 45
19
 
20
  ## Model description
21
 
@@ -35,58 +35,13 @@ More information needed
35
 
36
  The following hyperparameters were used during training:
37
  - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': 1e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
38
- - training_precision: float32
39
 
40
  ### Training results
41
 
42
  | Train Loss | Validation Loss | Epoch |
43
  |:----------:|:---------------:|:-----:|
44
- | 3.6637 | 2.2056 | 0 |
45
- | 2.0866 | 1.6338 | 1 |
46
- | 1.7565 | 1.4797 | 2 |
47
- | 1.5604 | 1.4183 | 3 |
48
- | 1.4554 | 1.2678 | 4 |
49
- | 1.3784 | 1.2015 | 5 |
50
- | 1.2908 | 0.9909 | 6 |
51
- | 1.2265 | 1.1029 | 7 |
52
- | 1.2524 | 1.0225 | 8 |
53
- | 1.1722 | 1.1167 | 9 |
54
- | 1.1396 | 0.9434 | 10 |
55
- | 1.1002 | 1.0683 | 11 |
56
- | 1.0693 | 1.0637 | 12 |
57
- | 1.0896 | 0.9698 | 13 |
58
- | 1.0590 | 0.9038 | 14 |
59
- | 1.0357 | 0.9302 | 15 |
60
- | 1.0557 | 0.8600 | 16 |
61
- | 1.0036 | 0.7892 | 17 |
62
- | 1.0283 | 0.9425 | 18 |
63
- | 0.9883 | 0.7521 | 19 |
64
- | 0.9797 | 0.7950 | 20 |
65
- | 0.9511 | 0.8072 | 21 |
66
- | 0.9023 | 0.8780 | 22 |
67
- | 0.9074 | 0.8745 | 23 |
68
- | 0.9324 | 0.7436 | 24 |
69
- | 0.8921 | 0.9032 | 25 |
70
- | 0.9098 | 0.8011 | 26 |
71
- | 0.8843 | 0.8527 | 27 |
72
- | 0.8756 | 0.7803 | 28 |
73
- | 0.8759 | 0.8922 | 29 |
74
- | 0.8472 | 0.8286 | 30 |
75
- | 0.8156 | 0.7801 | 31 |
76
- | 0.8401 | 0.7904 | 32 |
77
- | 0.8400 | 0.9007 | 33 |
78
- | 0.8368 | 0.6959 | 34 |
79
- | 0.8429 | 0.6646 | 35 |
80
- | 0.8496 | 0.7386 | 36 |
81
- | 0.8168 | 0.7544 | 37 |
82
- | 0.7927 | 0.8467 | 38 |
83
- | 0.8025 | 0.7375 | 39 |
84
- | 0.7893 | 0.7091 | 40 |
85
- | 0.7762 | 0.6758 | 41 |
86
- | 0.7516 | 0.8641 | 42 |
87
- | 0.7645 | 0.7587 | 43 |
88
- | 0.7790 | 0.7386 | 44 |
89
- | 0.7393 | 0.7801 | 45 |
90
 
91
 
92
  ### Framework versions
 
13
 
14
  This model is a fine-tuned version of [klue/roberta-base](https://huggingface.co/klue/roberta-base) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
+ - Train Loss: 3.4242
17
+ - Validation Loss: 2.0873
18
+ - Epoch: 0
19
 
20
  ## Model description
21
 
 
35
 
36
  The following hyperparameters were used during training:
37
  - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': 1e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
38
+ - training_precision: mixed_bfloat16
39
 
40
  ### Training results
41
 
42
  | Train Loss | Validation Loss | Epoch |
43
  |:----------:|:---------------:|:-----:|
44
+ | 3.4242 | 2.0873 | 0 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45
 
46
 
47
  ### Framework versions
tf_model.h5 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6cc3afe97f9bafd31730f2b742968b5ad338cea94f81a3ab4d7f4d0a69ad2be2
3
  size 542778024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ce76361ea503a4329dd9dd587798374252ce3b3ec817c457453af04b8e9fbf7
3
  size 542778024