Combining Vision Model with LLM: AI Visualizes Personal Memories

Merzmensch · x · 2026-08-13

The author continues to share updates on their #MERZmory project, where they previously trained a visual generative model (GAN) on 3,000 personal photos.

The author states that they could recognize real events from their life within the new images generated by the AI, and the results worked perfectly. They then combined this visual model with an LLM, also trained on their works, to further explore AI's interpretation of personal memory and experience.

Related event: Developer Trains GAN with Personal Photos to Reconstruct Memories(2 posts)→

Original post →

More from Fun

Fun channel →