1.8 KiB
Jibo Workshop HRI 2024
Resources:
https://github.com/jibodev/ShofEL2-for-T124-Jibo/tree/t124 - Github Repo https://hri2024.jibo.media.mit.edu/Setting-Up-Your-Codespace - Guide to setup a repo
About:
From the jibo workshop i have discovered that they have actually included a lot of info on how jibo expects the server to give it data, and by that i dont mean that it uses ESML, but what is available to use in every response!. Furthermore we also have access to some "examples" through the git repo on how to setup a Robot Os container, and I was able to extract the source code from the scratch extension over at Scratch Playground.
Were gonna have to analyse the stuff to go further one , and as i was looking at the files i discovered a file named jibo_chatGPT.ipynb which sparked some attention, and it appears that they have made a honest attempt at trying to hook it up to chat gpt
And there is a section where they clearly tried to make it act like jibo
def main(args=None):
teleop_connection = JiboTeleop()
time.sleep(2)
system_message = "You are Jibo, a friendly social robot who likes to chat about different topics." \
"Pick a random topic to discuss and use personalized stories." \
"Always respond in one short fun and exciting sentence, " \
"and use language that is easy to understand for all ages." \
" Give turns to user to respond."
# Spin in a separate thread
thread = threading.Thread(target=rclpy.spin, args=(teleop_connection, ), daemon=True)
thread.start()
...