Multitask Indonesian Sentiment & Sarcasm Analysis

This repository contains the fine-tuned checkpoint of IndoBERTweet trained under a multi-task learning configuration to predict both Sentiment (negative, neutral, positive, mixed) and Sarcasm (non_sarcastic, sarcastic).

Dataset

The model was fine-tuned on the Indonesian Reddit sarcasm dataset jwnt4/reddit-id-sarcasm-cleaned.

How to Load

Use the custom model.py module included in this repository to instantiate the architecture, and load the weights via PyTorch state_dict:

import torch
from model import SarcasmAwareSentimentModel, SarcasmAwareSentimentConfig

config = SarcasmAwareSentimentConfig()
model = SarcasmAwareSentimentModel(config)
model.load_state_dict(torch.load("best_model.pt", map_location="cpu"))
Downloads last month
40
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support