hanifnoerr
commited on
Commit
•
71de300
1
Parent(s):
2704cda
Upload README.md
Browse files
README.md
ADDED
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- indonlu
|
7 |
+
metrics:
|
8 |
+
- accuracy
|
9 |
+
- f1
|
10 |
+
model-index:
|
11 |
+
- name: Fine-tuned-Indonesian-Sentiment-Classifier
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Text Classification
|
15 |
+
type: text-classification
|
16 |
+
dataset:
|
17 |
+
name: indonlu
|
18 |
+
type: indonlu
|
19 |
+
config: smsa
|
20 |
+
split: validation
|
21 |
+
args: smsa
|
22 |
+
metrics:
|
23 |
+
- name: Accuracy
|
24 |
+
type: accuracy
|
25 |
+
value: 0.9317460317460318
|
26 |
+
- name: F1
|
27 |
+
type: f1
|
28 |
+
value: 0.9034223843742829
|
29 |
+
language:
|
30 |
+
- id
|
31 |
+
pipeline_tag: text-classification
|
32 |
+
|
33 |
+
widget:
|
34 |
+
- text: "Kalo kamu WFH emang kerja?"
|
35 |
+
- text: "dia tidak cantik"
|
36 |
+
---
|
37 |
+
|
38 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
39 |
+
should probably proofread and complete it, then remove this comment. -->
|
40 |
+
|
41 |
+
# Fine-tuned-Indonesian-Sentiment-Classifier
|
42 |
+
|
43 |
+
This model is a fine-tuned version of [indobenchmark/indobert-base-p1](https://huggingface.co/indobenchmark/indobert-base-p1) on the [IndoNLU's SmSA](https://huggingface.co/datasets/indonlp/indonlu) dataset.
|
44 |
+
It achieves the following results on the evaluation dataset:
|
45 |
+
- Loss: 0.3233
|
46 |
+
- Accuracy: 0.9317
|
47 |
+
- F1: 0.9034
|
48 |
+
|
49 |
+
And the results of the test dataset:
|
50 |
+
- Accuracy: 0.928
|
51 |
+
- F1 macro: 0.9113470780757361
|
52 |
+
- F1 micro: 0.928
|
53 |
+
- F1 weighted: 0.9261959965604815
|
54 |
+
|
55 |
+
## Model description
|
56 |
+
|
57 |
+
This model can be used to determine the sentiment of a text with three possible outputs [positive, negative, or neutral]
|
58 |
+
|
59 |
+
## How to use
|
60 |
+
|
61 |
+
```python
|
62 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
63 |
+
|
64 |
+
Pre-trained = "hanifnoerr/Fine-tuned-Indonesian-Sentiment-Classifier"
|
65 |
+
tokenizer = AutoTokenizer.from_pretrained(Pre-trained)
|
66 |
+
model = AutoModelForSequenceClassification.from_pretrained(Pre-trained)
|
67 |
+
```
|
68 |
+
### make classification
|
69 |
+
```python
|
70 |
+
pretrained_name = "hanifnoerr/Fine-tuned-Indonesian-Sentiment-Classifier"
|
71 |
+
class_model = pipeline(tokenizer=pretrained_name, model=pretrained_name)
|
72 |
+
|
73 |
+
test_data = "buku ini jelek sekali"
|
74 |
+
class_model(test_data)
|
75 |
+
```
|
76 |
+
output: [{'label': 'negative', 'score': 0.9996247291564941}]
|
77 |
+
|
78 |
+
### Training results
|
79 |
+
|
80 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|
81 |
+
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:------:|
|
82 |
+
| 0.08 | 1.0 | 688 | 0.3532 | 0.9310 | 0.9053 |
|
83 |
+
| 0.0523 | 2.0 | 1376 | 0.3233 | 0.9317 | 0.9034 |
|
84 |
+
| 0.045 | 3.0 | 2064 | 0.3949 | 0.9286 | 0.8995 |
|
85 |
+
| 0.0252 | 4.0 | 2752 | 0.4662 | 0.9310 | 0.9049 |
|
86 |
+
| 0.0149 | 5.0 | 3440 | 0.6251 | 0.9246 | 0.8899 |
|
87 |
+
| 0.0091 | 6.0 | 4128 | 0.6148 | 0.9254 | 0.8928 |
|
88 |
+
| 0.0111 | 7.0 | 4816 | 0.6259 | 0.9222 | 0.8902 |
|
89 |
+
| 0.0106 | 8.0 | 5504 | 0.6123 | 0.9238 | 0.8882 |
|
90 |
+
| 0.0092 | 9.0 | 6192 | 0.6353 | 0.9230 | 0.8928 |
|
91 |
+
| 0.0085 | 10.0 | 6880 | 0.6733 | 0.9254 | 0.8989 |
|
92 |
+
| 0.0062 | 11.0 | 7568 | 0.6666 | 0.9302 | 0.9027 |
|
93 |
+
| 0.0036 | 12.0 | 8256 | 0.7578 | 0.9230 | 0.8962 |
|
94 |
+
| 0.0055 | 13.0 | 8944 | 0.7378 | 0.9270 | 0.8947 |
|
95 |
+
| 0.0023 | 14.0 | 9632 | 0.7758 | 0.9230 | 0.8978 |
|
96 |
+
| 0.0009 | 15.0 | 10320 | 0.7051 | 0.9278 | 0.9006 |
|
97 |
+
| 0.0033 | 16.0 | 11008 | 0.7442 | 0.9214 | 0.8902 |
|
98 |
+
| 0.0 | 17.0 | 11696 | 0.7513 | 0.9254 | 0.8974 |
|
99 |
+
| 0.0 | 18.0 | 12384 | 0.7554 | 0.9270 | 0.8999 |
|
100 |
+
|
101 |
+
Although trained with 18 epochs, this model uses the best weight (Epoch 2)
|
102 |
+
|
103 |
+
### Framework versions
|
104 |
+
|
105 |
+
- Transformers 4.27.4
|
106 |
+
- Pytorch 2.0.0+cu118
|
107 |
+
- Datasets 2.11.0
|
108 |
+
- Tokenizers 0.13.3
|