Hugging Face's logo

Noodle-bg
/
BERT_Text_classification_noisy_FULL

Text Classification
Transformers
Safetensors
bert
Generated from Trainer
Model card Files
xet
 Community
Noodle-bg commited on
Commit
4aa2c05
·
verified ·
1 Parent(s): 0727576

Model save

Browse files
Files changed (1) hide show
  1. README.md +87 -0
README.md ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: bert-base-uncased
4
+ tags:
5
+ - generated_from_trainer
6
+ metrics:
7
+ - accuracy
8
+ - f1
9
+ - precision
10
+ - recall
11
+ model-index:
12
+ - name: BERT_Text_classification_noisy_FULL
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # BERT_Text_classification_noisy_FULL
20
+
21
+ This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on an unknown dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.5701
24
+ - Accuracy: 0.8552
25
+ - F1: 0.8469
26
+ - Precision: 0.8499
27
+ - Recall: 0.8466
28
+
29
+ ## Model description
30
+
31
+ More information needed
32
+
33
+ ## Intended uses & limitations
34
+
35
+ More information needed
36
+
37
+ ## Training and evaluation data
38
+
39
+ More information needed
40
+
41
+ ## Training procedure
42
+
43
+ ### Training hyperparameters
44
+
45
+ The following hyperparameters were used during training:
46
+ - learning_rate: 5e-05
47
+ - train_batch_size: 16
48
+ - eval_batch_size: 32
49
+ - seed: 42
50
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
51
+ - lr_scheduler_type: linear
52
+ - lr_scheduler_warmup_steps: 100
53
+ - num_epochs: 3
54
+
55
+ ### Training results
56
+
57
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
58
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
59
+ | 2.8254 | 0.14 | 100 | 2.1467 | 0.4831 | 0.4014 | 0.4638 | 0.4628 |
60
+ | 1.7026 | 0.28 | 200 | 1.0972 | 0.7123 | 0.6713 | 0.6981 | 0.6939 |
61
+ | 1.207 | 0.42 | 300 | 0.9107 | 0.7524 | 0.7200 | 0.7407 | 0.7315 |
62
+ | 1.0468 | 0.56 | 400 | 0.7801 | 0.7861 | 0.7586 | 0.7829 | 0.7676 |
63
+ | 1.0953 | 0.71 | 500 | 0.8004 | 0.7852 | 0.7649 | 0.7778 | 0.7685 |
64
+ | 1.0857 | 0.85 | 600 | 0.8224 | 0.7678 | 0.7367 | 0.7775 | 0.7507 |
65
+ | 0.9787 | 0.99 | 700 | 0.6742 | 0.8111 | 0.7968 | 0.8069 | 0.7976 |
66
+ | 0.7944 | 1.13 | 800 | 0.6627 | 0.8222 | 0.8068 | 0.8162 | 0.8077 |
67
+ | 0.8078 | 1.27 | 900 | 0.6222 | 0.8306 | 0.8194 | 0.8260 | 0.8187 |
68
+ | 0.8101 | 1.41 | 1000 | 0.6620 | 0.8279 | 0.8158 | 0.8276 | 0.8155 |
69
+ | 0.7668 | 1.55 | 1100 | 0.5905 | 0.8439 | 0.8361 | 0.8400 | 0.8356 |
70
+ | 0.7133 | 1.69 | 1200 | 0.6085 | 0.8328 | 0.8231 | 0.8297 | 0.8225 |
71
+ | 0.7621 | 1.84 | 1300 | 0.6039 | 0.8375 | 0.8293 | 0.8368 | 0.8283 |
72
+ | 0.6461 | 1.98 | 1400 | 0.6176 | 0.8392 | 0.8298 | 0.8356 | 0.8292 |
73
+ | 0.5791 | 2.12 | 1500 | 0.5925 | 0.8524 | 0.8439 | 0.8497 | 0.8440 |
74
+ | 0.602 | 2.26 | 1600 | 0.5877 | 0.8473 | 0.8389 | 0.8441 | 0.8382 |
75
+ | 0.5459 | 2.4 | 1700 | 0.5888 | 0.8534 | 0.8460 | 0.8528 | 0.8459 |
76
+ | 0.569 | 2.54 | 1800 | 0.6118 | 0.8467 | 0.8393 | 0.8448 | 0.8387 |
77
+ | 0.5138 | 2.68 | 1900 | 0.5797 | 0.8561 | 0.8474 | 0.8520 | 0.8474 |
78
+ | 0.5427 | 2.82 | 2000 | 0.5678 | 0.8578 | 0.8498 | 0.8531 | 0.8493 |
79
+ | 0.5437 | 2.97 | 2100 | 0.5701 | 0.8552 | 0.8469 | 0.8499 | 0.8466 |
80
+
81
+
82
+ ### Framework versions
83
+
84
+ - Transformers 4.36.2
85
+ - Pytorch 2.1.2+cu121
86
+ - Datasets 2.16.1
87
+ - Tokenizers 0.15.2