README.md

---
title: CV InsectClassifier # Ini dari bagian remote, kamu bisa ubah namanya jadi lebih deskriptif kalau mau
emoji: 🐞 # Kamu bisa pilih emoji yang kamu suka, misal tetap 🐞
colorFrom: gray
colorTo: green
sdk: docker
pinned: false
---

# 🐞 Insect Classifier - FastAPI

A web-based insect classification app using two TensorFlow models (CNN and MobileNet), deployed with FastAPI. Users can upload insect images and receive predictions from both models, along with accuracy and descriptions.

## 🧠 Models
- CNN model (`ProyekCV_model.h5`)
- MobileNet model (`ProyekCV_model_v2.h5`)

## ⚙️ Tech Stack
- FastAPI
- TensorFlow & Keras
- HTML/CSS (for frontend)
- Uvicorn (as ASGI server)
- Python
- NumPy
# - Pandas # Kamu tidak pakai Pandas di main.py, sebaiknya dihapus dari daftar
# - Matplotlib & Seaborn # Kamu tidak pakai ini untuk runtime aplikasi, sebaiknya dihapus atau pindah ke "Development Dependencies"
# - Scikit-learn # Kamu tidak pakai ini, sebaiknya dihapus

## 📦 Dataset

This project uses the [Insects Recognition Dataset](https://www.kaggle.com/datasets/hammaadali/insects-recognition) by Hammaad Ali, available on Kaggle.

**Dataset Features:**
- Contains high-quality images of 5 different insect classes. There are grasshopper, butterfly, mosquito, ladybird and dragonfly.
- Organized into labeled folders for each class.
- Ideal for supervised image classification tasks.
- Image format: `.jpg`

The dataset was used to train both the CNN and MobileNet models included in this project.

## 🚀 How to Run Locally

Make sure you have all dependencies installed and your virtual environment activated.

**Step 1: Start the FastAPI backend**
Open a terminal and run:

```bash
venv\Scripts\activate
uvicorn main:app --reload --host 0.0.0.0 --port 8000 # <-- PERBAIKI INI: sesuaikan dengan command lokal mu
```