Hugging Face's logo

3728km
/
fined-tune-thai-sentiment

Text Classification
Transformers
TensorBoard
Safetensors
camembert
Generated from Trainer
Model card Files
xet
Metrics Community
3728km commited on
Commit
6237aef
·
verified ·
1 Parent(s): 94ca8b8

End of training

Browse files
Files changed (2) hide show
  1. README.md +19 -29
  2. model.safetensors +1 -1
README.md CHANGED
@@ -19,11 +19,11 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  This model is a fine-tuned version of [airesearch/wangchanberta-base-att-spm-uncased](https://huggingface.co/airesearch/wangchanberta-base-att-spm-uncased) on an unknown dataset.
21
  It achieves the following results on the evaluation set:
22
- - Loss: 0.6408
23
- - Accuracy: 0.9181
24
- - F1-score: 0.9166
25
- - Precision: 0.9158
26
- - Recall: 0.9181
27
 
28
  ## Model description
29
 
@@ -42,44 +42,34 @@ More information needed
42
  ### Training hyperparameters
43
 
44
  The following hyperparameters were used during training:
45
- - learning_rate: 2e-05
46
  - train_batch_size: 8
47
  - eval_batch_size: 8
48
  - seed: 42
49
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
  - lr_scheduler_type: linear
51
- - lr_scheduler_warmup_steps: 680
52
- - num_epochs: 20
53
 
54
  ### Training results
55
 
56
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1-score | Precision | Recall |
57
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:---------:|:------:|
58
- | 0.0565 | 1.0 | 85 | 0.8049 | 0.8772 | 0.8749 | 0.8732 | 0.8772 |
59
- | 0.0376 | 2.0 | 170 | 0.8149 | 0.8713 | 0.8712 | 0.8711 | 0.8713 |
60
- | 0.0614 | 3.0 | 255 | 0.7712 | 0.8772 | 0.8749 | 0.8732 | 0.8772 |
61
- | 0.0391 | 4.0 | 340 | 0.8746 | 0.8713 | 0.8735 | 0.8762 | 0.8713 |
62
- | 0.0306 | 5.0 | 425 | 0.8107 | 0.8889 | 0.8847 | 0.8839 | 0.8889 |
63
- | 0.0245 | 6.0 | 510 | 0.8521 | 0.9006 | 0.8963 | 0.8961 | 0.9006 |
64
- | 0.0387 | 7.0 | 595 | 0.7422 | 0.9123 | 0.9103 | 0.9091 | 0.9123 |
65
- | 0.055 | 8.0 | 680 | 0.7637 | 0.9123 | 0.9077 | 0.9065 | 0.9123 |
66
- | 0.0598 | 9.0 | 765 | 0.6920 | 0.9181 | 0.9159 | 0.9146 | 0.9181 |
67
- | 0.1463 | 10.0 | 850 | 0.6993 | 0.9181 | 0.9151 | 0.9142 | 0.9181 |
68
- | 0.1465 | 11.0 | 935 | 0.7118 | 0.8947 | 0.8955 | 0.8967 | 0.8947 |
69
- | 0.1595 | 12.0 | 1020 | 0.7225 | 0.8947 | 0.8981 | 0.9029 | 0.8947 |
70
- | 0.1518 | 13.0 | 1105 | 0.6408 | 0.9181 | 0.9166 | 0.9158 | 0.9181 |
71
- | 0.1437 | 14.0 | 1190 | 0.7732 | 0.8889 | 0.8909 | 0.8934 | 0.8889 |
72
- | 0.0825 | 15.0 | 1275 | 0.7639 | 0.9006 | 0.8997 | 0.8991 | 0.9006 |
73
- | 0.023 | 16.0 | 1360 | 0.7652 | 0.9064 | 0.9056 | 0.9050 | 0.9064 |
74
- | 0.0614 | 17.0 | 1445 | 0.7828 | 0.8947 | 0.8955 | 0.8967 | 0.8947 |
75
- | 0.0394 | 18.0 | 1530 | 0.8096 | 0.8947 | 0.8938 | 0.8932 | 0.8947 |
76
- | 0.0678 | 19.0 | 1615 | 0.8152 | 0.8947 | 0.8938 | 0.8932 | 0.8947 |
77
- | 0.0749 | 20.0 | 1700 | 0.8122 | 0.8947 | 0.8938 | 0.8932 | 0.8947 |
78
 
79
 
80
  ### Framework versions
81
 
82
- - Transformers 4.50.3
83
  - Pytorch 2.6.0+cu124
84
  - Datasets 3.5.0
85
  - Tokenizers 0.21.1
 
19
 
20
  This model is a fine-tuned version of [airesearch/wangchanberta-base-att-spm-uncased](https://huggingface.co/airesearch/wangchanberta-base-att-spm-uncased) on an unknown dataset.
21
  It achieves the following results on the evaluation set:
22
+ - Loss: 0.3544
23
+ - Accuracy: 0.9282
24
+ - F1-score: 0.9278
25
+ - Precision: 0.9276
26
+ - Recall: 0.9282
27
 
28
  ## Model description
29
 
 
42
  ### Training hyperparameters
43
 
44
  The following hyperparameters were used during training:
45
+ - learning_rate: 5e-05
46
  - train_batch_size: 8
47
  - eval_batch_size: 8
48
  - seed: 42
49
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
  - lr_scheduler_type: linear
51
+ - lr_scheduler_warmup_steps: 181
52
+ - num_epochs: 10
53
 
54
  ### Training results
55
 
56
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1-score | Precision | Recall |
57
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:---------:|:------:|
58
+ | 0.8746 | 1.0 | 91 | 0.8613 | 0.6133 | 0.4662 | 0.3761 | 0.6133 |
59
+ | 0.8086 | 2.0 | 182 | 0.8758 | 0.5746 | 0.4955 | 0.4768 | 0.5746 |
60
+ | 0.9223 | 3.0 | 273 | 0.9218 | 0.6133 | 0.4662 | 0.3761 | 0.6133 |
61
+ | 0.8561 | 4.0 | 364 | 0.7430 | 0.6630 | 0.5899 | 0.6325 | 0.6630 |
62
+ | 0.6694 | 5.0 | 455 | 0.5335 | 0.7845 | 0.7507 | 0.7289 | 0.7845 |
63
+ | 0.5792 | 6.0 | 546 | 0.4365 | 0.8287 | 0.8227 | 0.8239 | 0.8287 |
64
+ | 0.3046 | 7.0 | 637 | 0.4033 | 0.8840 | 0.8834 | 0.8930 | 0.8840 |
65
+ | 0.2004 | 8.0 | 728 | 0.3544 | 0.9282 | 0.9278 | 0.9276 | 0.9282 |
66
+ | 0.1443 | 9.0 | 819 | 0.4025 | 0.9171 | 0.9180 | 0.9199 | 0.9171 |
67
+ | 0.0765 | 10.0 | 910 | 0.4116 | 0.9227 | 0.9238 | 0.9269 | 0.9227 |
 
 
 
 
 
 
 
 
 
 
68
 
69
 
70
  ### Framework versions
71
 
72
+ - Transformers 4.51.3
73
  - Pytorch 2.6.0+cu124
74
  - Datasets 3.5.0
75
  - Tokenizers 0.21.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:be82034ecf6a01767275be153d9c3c352c3f605ca5ad002918e263490a6435f7
3
  size 421011004
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ac4cd541c82353ec2d373f0318aaf08454a58987defa21b0131c7dd76739930
3
  size 421011004