AIbase

Robotic-Action-Frame-Prediction-with-InstructPix2Pix

Public

This repository contains the code and configuration files for training a multimodal fine-tuned `InstructPix2Pix` model to predict future robotic action frames. The model generates 256×256 resolution images conditioned on a current observation and textual instruction

Creat2025-04-29T15:07:58
Update2025-05-15T06:49:01
6
Stars
0
Stars Increase