Prismatic VLM REPL

From Humanoid Robots Wiki
Revision as of 23:00, 20 June 2024 by Vrtnis (talk | contribs)
Jump to: navigation, search

The K-Scale OpenVLA adaptation by User:Paweł is at https://github.com/kscalelabs/openvla

Here are some suggestions to running the generate.py REPL Script from the repo if you just want to try out OpenVLA.

Prerequisites

Before running the script, ensure you have the following:

  • Python 3.8 or higher installed
  • NVIDIA GPU with CUDA support (optional but recommended for faster processing)
  • Hugging Face account and token for accessing certain models

Setting Up the Environment

In addition to installing requirements-min.txt from the repo, you probably need to install rich, tensorflow_graphics, tensorflow-datasets and dlimp.