A high-performance Computer Vision API built to identify recyclable waste types. This project uses FastAPI to serve a Google ViT (Vision Transformer) model that has been fine-tuned using LoRA (Low-Rank Adaptation) to recognize 15 specific categories of waste.🚀

• ** State-of-the-art Vision **: Uses google/vit-base-patch16-224 as the backbone.
• ** Efficient Fine-Tuning ** : Uses PEFT & LoRA adapters (~10MB) - instead of full model weights (~350MB).
• ** Fast Inference: ** Optimized for local CPU execution.
• ** Async API ** : Built on FastAPI for non-blocking performance.
• ** 30 Classes ** : Specifically trained to detect common household recyclables.

• ** Framework ** : FastAPI (Python)
• ** ML Engine ** : PyTorch
• ** Model Library ** : Hugging Face Transformers & PEFT
• ** Image Processing ** : Pillow (PIL)
• ** Server ** : Uvicorn📂