Starting an LLM inference engine from scratch. First stop: reading every line of Karpathy's llama2.c. It's 500 lines of C that runs a transformer. This is the vibe.
What if a phone could replace a $50K LiDAR rig?
wayfr turns a walkthrough video into a fully annotated 3D scene: object detection, segmentation, world coords.
Same space, different Personas for each team.
@ansh_is_g@Aditya130805
https://t.co/BQTHLo7lWG
@ansh_is_g@Aditya130805 Here's how it works:
1. A contributor records a walkthrough video
2. Our pipeline (MapAnything) reconstructs the space in 3D, then Grounding DINO + SAM2 to detect and segment every object
3. 2D detections are lifted into world coordinates, a fully labeled, explorable 3D scene
We won HackIllinois (@supermemory’s and @stripe’s 2nd place prizes)!
We built Superstyle, a way to see how your clothes fit before you buy them with hyper realistic images and clothing website integrations.
Built it with @IFirosh and @xiezilla
Devpost: https://t.co/tQVWBIdYpQ