radityapranata
commited on
Commit
•
b6edf72
1
Parent(s):
7aeb00c
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- id
|
4 |
+
---
|
5 |
+
|
6 |
+
UPDATE V3 DENGAN MENGGUNAKAN 12512 DATASET PENGUJIAN
|
7 |
+
# Model Card radityapranata/absabert-keluhanpln-v5
|
8 |
+
update from radityapranata/absabert-keluhanpln-v4
|
9 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
10 |
+
Basic Model ABSA BERT KELUHAN PLN dalam bahasa indonesia merupakan model yang dihasilkan dari : ##bert-base-uncased
|
11 |
+
[this raw template](https://huggingface.co/bert-base-uncased).
|
12 |
+
|
13 |
+
|
14 |
+
## Model Description
|
15 |
+
<!-- Provide a longer summary of what this model is. -->
|
16 |
+
Dataset yang digunakan disini merupakan kumpulan data keluhan pelanggan
|
17 |
+
dalam bahasa indonesia pada PLN Mobile.
|
18 |
+
|
19 |
+
Dengan pengukuran aspect yang terbagi meliputi :
|
20 |
+
"Layanan Pelanggan" , "Ketersediaan Produk" , "Kebijakan Usaha" , "Pemulihan Layanan".
|
21 |
+
|
22 |
+
Hal ini mengutip dari jenis pengukuran yag dilakukan oleh Simon J. Bell di tahun 2006
|
23 |
+
dari University of Melbourne dalam publikasinya yang berjudul
|
24 |
+
Coping With Customer Complaints
|
25 |
+
|
26 |
+
### How to use
|
27 |
+
|
28 |
+
You can use this model directly with a pipeline for masked language modeling:
|
29 |
+
|
30 |
+
```python
|
31 |
+
>>>pip install torch
|
32 |
+
>>>pip install transformers
|
33 |
+
>>>from transformers import BertTokenizer, BertForSequenceClassification
|
34 |
+
>>>import torch
|
35 |
+
|
36 |
+
model_name = "radityapranata/absabert-keluhanpln-v3"
|
37 |
+
tokenizer = BertTokenizer.from_pretrained(model_name)
|
38 |
+
model = BertForSequenceClassification.from_pretrained(model_name)
|
39 |
+
absa_pipeline = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
|
40 |
+
|
41 |
+
text = "informasi tambah daya" #"Token listrik tidak dapat terisi, kwh meter tulisan periksa."
|
42 |
+
inputs = tokenizer(text, return_tensors="pt")
|
43 |
+
result = absa_pipeline(text)
|
44 |
+
|
45 |
+
outputs = model(**inputs)
|
46 |
+
logits = outputs.logits
|
47 |
+
|
48 |
+
get_aspect = torch.argmax(logits, dim=1).item()
|
49 |
+
aspects = ["Layanan Pelanggan", "Ketersediaan Produk", "Kebijakan Usaha", "Pemulihan Layanan"]
|
50 |
+
aspect = aspects[get_aspect]
|
51 |
+
|
52 |
+
for aspect_result in result:
|
53 |
+
Sentiment = aspect_result["label"]
|
54 |
+
Score = aspect_result["score"]
|
55 |
+
print("Aspect:", aspect)
|
56 |
+
print(f"Sentiment: {Sentiment}, Score: {Score}")
|
57 |
+
```
|