from transformers import BertTokenizer, BertForSequenceClassification
import torch

# Load model and tokenizer
tokenizer = BertTokenizer.from_pretrained("MrPanuwit/thai-youtube-sentiment-subtask1")
model = BertForSequenceClassification.from_pretrained("MrPanuwit/thai-youtube-sentiment-subtask1")

# Example usage
text = "วิดีโอนี้สนุกมาก"
inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=128)

with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.argmax(outputs.logits, dim=-1)
    
print(f"Prediction: {predictions.item()}")

Training Details

Framework: PyTorch + Transformers
Optimizer: AdamW
Max Sequence Length: 128

Limitations

Trained specifically on Thai YouTube comments
May not perform well on other text domains
Limited to the specific sentiment categories in training data

Citation

@misc{thai-youtube-sentiment,
  author = {Your Name},
  title = {Thai YouTube Comment Sentiment Analysis},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/MrPanuwit/thai-youtube-sentiment-subtask1}
}