Skip to content

Latest commit

 

History

History
5 lines (3 loc) · 369 Bytes

README.md

File metadata and controls

5 lines (3 loc) · 369 Bytes

Unofficial Implementation of CLIP: Learning Transferable Visual Models From Natural Language Supervision paper

The clip train.pynb file contains the code to train a simple CLIP model for zero-shot recognition.

I used this Github Repo - [https://github.com/moein-shariatnia/OpenAI-CLIP] - as reference to understand its implementation from scratch in Pytorch.