metadata
language: vi
tags:
- spam-detection
- vietnamese
- transformer
license: apache-2.0
datasets:
- visolex/ViSpamReviews
metrics:
- accuracy
- f1
model-index:
- name: visobert-spam-classification
results:
- task:
type: text-classification
name: Spam Detection (Multi-Class)
dataset:
name: ViSpamReviews
type: custom
metrics:
- name: Accuracy
type: accuracy
value: <INSERT_ACCURACY>
- name: F1 Score
type: f1
value: <INSERT_F1_SCORE>
base_model:
- uitnlp/visobert
pipeline_tag: text-classification
ViSoBERT-Spam-MultiClass
Fine-tuned from uitnlp/visobert on ViSpamReviews for multi-class spam classification.
Task: 4-way classification (
SpamLabel: 0=NO-SPAM, 1=SPAM-1, 2=SPAM-2, 3=SPAM-3)Dataset: ViSpamReviews
Hyperparameters
- Batch size: 32
- LR: 3e-5
- Epochs: 100
- Max seq len: 256
Usage
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("visolex/visobert-spam-classification")
model = AutoModelForSequenceClassification.from_pretrained("visolex/visobert-spam-classification")
text = "Chỉ nói về thương hiệu thôi."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=256)
pred = model(**inputs).logits.argmax(dim=-1).item()
label_map = {0: "NO-SPAM",1: "SPAM-1",2: "SPAM-2",3: "SPAM-3"}
print(label_map[pred])