Flow Matching for Generative Modeling: https://buff.ly/3T4bb9W
Latent Consistency Models: https://buff.ly/3COzdmh

Latent Consistency Models (LCM): While FM rewrites the rules, LCM focuses on optimizing the existing technique. By “teaching” a smaller, faster model to mimic diffusion models, LCM drastically reduces the time needed for inference. We go from 100s of denoising step to 1-4.

Flow Matching (FM): A reimagination of how image generation could work. Instead of step-by-step denoising, FM replaces it with a generalized approach that’s faster, more flexible, and requires fewer steps to generate an image. It's a new ground-up approach to image generation.

The moment a new capability emerges like diffusion-based image generation (e.g. StableDiffusion), we see that human creativity takes off. Our imagination runs wild and we dream of applications that might seem almost magical.

But then we hit some roadblocks 🧵

4/ In the second attempt, I added four small markers to the image—highlighting key spots like cars and mountains! GPT-4o nailed it.
Not only did it correctly identify the location as Albania, but it also looked at number plates and suggested the exact area I was heading to!

3/ First, I uploaded the photo as it was.
The AI’s response was vague and unhelpful. No clear idea where the picture was taken.