A simple open-sourced SigLIP model finetuned on Genshin Impact's image-text pairs.

Pocket-Sized Multimodal AI for content understanding and generation across multilingual texts, images, and 🔜 video, up to 5x faster than OpenAI CLIP and LLaVA 🖼️ & 🖋️

Clipora is a powerful toolkit for fine-tuning OpenCLIP models using Low Rank Adapters (LoRA).

Using Docker compose to start up Triton with the OpenClip model, to encode text in to vectors

Text-to-image search with OpenCLIP, Docker, Flask, Faiss, etc. and a basic front-end.

[Official] [IROS 2024] A goal-oriented planning to lift VLN performance for Closed-Loop Navigation: Simple, Yet Effective

A quick test using Supabase to test if pgvector is the right solution for a different project. The embedding model I will be using OpenAI's Clip Model which will allow me to embed images and text in the same vector space.

Searching Images: From Clip And Beyond

Reproducible scaling laws for contrastive language-image learning (https://arxiv.org/abs/2212.07143)

Group images by provided labels using OpenAI/CLIP

CLIP based Zero Shot Instance Segmentation

use SAM and OpenCLIP to perform zero-shot object detection using COCO 2017 val split.

Using Segment-Anything and CLIP to generate pixel-aligned semantic features.