File size: 4,883 Bytes
29f6ccc fd9cdee 29f6ccc 3ce4bfc de71074 fd9cdee 29f6ccc 1b858e0 64dcedf 1b858e0 64dcedf 1b858e0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 |
---
title: Image Classifier With CNN
emoji: 🚀
colorFrom: red
colorTo: red
sdk: docker
app_port: 8501
tags:
- streamlit
pinned: true
short_description: A CNN-based image classification with TensorFlow
license: mit
---
# 🖼️ Image Classification with ResNet50
## Table of Contents
1. [Project Description](#1-project-description)
2. [Methodology & Key Features](#2-methodology--key-features)
3. [Technology Stack](#3-technology-stack)
4. [Deployment Options](#4-deployment-options)
5. [Model](#5-model)
## 1. Project Description
This project implements an **image classification system** using a pre-trained **ResNet50 CNN** (trained on ImageNet with 1,000+ categories).
The repository includes **two versions**:
1. **Full Architecture (local, Dockerized):**
- Built as a **multi-service ML system** with FastAPI (backend inference), Redis (queue/cache), and Streamlit (frontend).
- Designed to demonstrate a scalable, production-like workflow.
2. **Lightweight Streamlit App (deployed):**
- Adapted into a **single Streamlit app** for cost-effective deployment on **Hugging Face Spaces**.
- Lets users upload an image and instantly see the **predicted category** with confidence.
This dual approach allows others to explore both a **realistic ML architecture** and a **lightweight, deployable demo**.
> [!IMPORTANT]
>
> - Check out the deployed app here: 👉️ [Image Classification App](https://huggingface.co/spaces/iBrokeTheCode/Image_Classifier_with_CNN) 👈️
> - Check out the source code for multi-service setup: 👉️ [Multi-Service - Source Code](https://huggingface.co/spaces/iBrokeTheCode/Image_Classifier_with_CNN/tree/main) 👈️
> - Check out the source code for lightweight Streamlit App: 👉️ [Lightweight App - Source Code](https://huggingface.co/spaces/iBrokeTheCode/Image_Classifier_with_CNN/tree/main/src) 👈️
## 2. Methodology & Key Features
- **Model:** ResNet50 (pre-trained on **ImageNet** with 1,000+ classes).
- **Pipeline:** Input images are resized, normalized, and passed to the model.
- **Output:** Top-1 prediction with **confidence score** is displayed.
- **Multi-service architecture:**
- **FastAPI** serves inference requests.
- **Redis** handles caching and task queueing.
- **Streamlit** provides the interactive UI.
- **Lightweight deployment:** Direct **Streamlit-only** version for Hugging Face Spaces.
## 3. Technology Stack
This project was built using the following technologies:
**Deployment & Hosting:**
- [Docker](https://www.docker.com/) – containerization for the full architecture.
- [Hugging Face Spaces](https://huggingface.co/docs/hub/spaces) – for lightweight deployment.
- [Streamlit](https://streamlit.io/) – interactive web app frontend.
**Backend & Infrastructure:**
- [FastAPI](https://fastapi.tiangolo.com/) – high-performance inference API.
- [Redis](https://redis.io/) – caching and message queue.
**Modeling & Training:**
- [TensorFlow / Keras](https://www.tensorflow.org/) – ResNet50 model (pre-trained on ImageNet).
**Development Tools:**
- [Ruff](https://github.com/charliermarsh/ruff) – Python linter and formatter.
- [uv](https://github.com/astral-sh/uv) – fast Python package installer and resolver.
## 4. Deployment Options
You can run this project in two ways:
### A. Run the Lightweight Version (Streamlit-only)
1. Clone the repo:
```bash
git clone https://huggingface.co/spaces/iBrokeTheCode/Image_Classifier_with_CNN
cd Image_Classifier_with_CNN
```
2. Install dependencies:
```bash
pip install -r requirements.txt
```
3. Run the app:
```bash
streamlit run src/streamlit_app.py
```
### B. Run the Full Architecture (Dockerized)
The repository also contains **Dockerfiles** for each service.
1. Clone the repo:
```bash
git clone https://huggingface.co/spaces/iBrokeTheCode/Image_Classifier_with_CNN
cd Image_Classifier_with_CNN
```
2. Pre-configure your environment variables:
```bash
cp .env.original .env
```
3. Create a network for containers
```bash
docker network create shared_network
```
4. Build and start all services:
```bash
docker compose up --build -d
# Stop services
docker compose down
```
5. Populate the database
```bash
cd api
cp .env.original .env
docker-compose up --build -d
```
6. Access the app at:
```
http://localhost:9090
```
Then use this credentials to pass the login:
- Username: admin@example.com
- Password: admin
7. Access the FastAPI app at:
```
http://localhost:8000/docs
```
## 5. Model
This project uses the pre-training model ResNet50.
- **Model:** [ResNet50](https://www.tensorflow.org/api_docs/python/tf/keras/applications/ResNet50)
- **Classes:** 1,000+ categories (objects, animals, everyday items).
|