The CLIP-MLOps-Pipeline is a Machine Learning Operations (MLOps) pipeline designed to train, evaluate, and deploy CLIP-based models for image-text processing. It includes modular components for data processing, model training, experiment tracking with MLflow, API deployment with FlaskAPI, and CI/CD automation using Jenkins and Docker.
Modular Design: Organized structure for models, data, training, and testing.
MLflow Tracking: Tracks model metrics and versioning.
Dockerized Deployment: API and training environments are containerized.
CI/CD Integration: Automated pipeline with Jenkins for testing and deployment.
FlaskAPI for Model Serving: Provides an efficient API for inference.
📦 CLIP-MLOps-Pipeline
├── 📂 src/
│ ├── 📂 models/
│ │ ├── image_encoder.py
│ │ ├── text_encoder.py
│ │ ├── transformer_encoder.py
│ │ ├── attn_head.py
│ │ ├── pos_embeds.py
│ │ ├── model.py
│ ├── 📂 data/
│ │ ├── dataset.py
│ │ ├── parameters.json
│ ├── 📂 training/
│ │ ├── train.py
│ │ ├── test.py
│ ├── 📂 api/
│ │ ├── main.py
│ │ ├── Dockerfile
│ │ ├── requirements.txt
│ ├── 📂 tests/
│ │ ├── test_model.py
│ │ ├── test_api.py
├── 📂 mlflow/
│ ├── tracking.py
├── 📂 docker/
│ ├── Dockerfile
├── 📂 ci-cd/
│ ├── Jenkinsfile
├── README.md
├── requirements.txt
├── .gitignore