--- title: Multilingual Emotion Recognition (EN / SI / TA) emoji: 😊 colorFrom: indigo colorTo: purple sdk: gradio sdk_version: 4.44.0 app_file: app.py python_version: 3.11 pinned: false --- # Multilingual Text Emotion Recognition App **English • සිංහල • தமிழ்** This is a simple web app that detects emotions from text in three languages: - **English** → Logistic Regression + TF-IDF (classical ML model) - **Sinhala** → Finetuned transformer model - **Tamil** → Finetuned transformer model ### Supported Languages & Models | Language | Model Type | Repository on Hugging Face | |------------|-------------------------|------------------------------------------------------------------| | English | TF-IDF + Logistic Regression | E-motionAssistant/Englsih_Trained_Model_LR | | Sinhala | Finetuned Transformer | E-motionAssistant/SInhala_Text_Emotion_Recognition_Model | | Tamil | Finetuned Transformer | E-motionAssistant/Tamil_Emotion_Recognition_Model | ### How to Use 1. Go to one of the tabs (English / සිංහල / தமிழ்) 2. Type or paste your text/message/tweet 3. The app will show the predicted emotion + confidence score ### Examples - English: "I hate mondays so much" → likely **sadness** or **anger** - Sinhala: "මම ගොඩක් දුකින් ඉන්නවා" → likely **sadness** - Tamil: "எனக்கு ரொம்ப கோபமா இருக்கு" → likely **anger** ### Technical Notes - Built with **Gradio** (v4.44.0) - Models are downloaded from Hugging Face at startup (using `hf_hub_download` and `pipeline`) - Uses **Python 3.11** to avoid compatibility issues - No GPU required — runs on CPU ### Want to run locally? ```bash git clone https://huggingface.co/spaces/YOUR_USERNAME/Space4 cd Space4 pip install -r requirements.txt python app.py