tianjunz
tianjunz
We used a random representation (e.g., projecting the state space using a random MLP network). It is very similar to the maze environment except changing the CNN network into a...
No, we don't pretrain/train the network. We use a random network for projecting. So pretty much a randomly initialized MLP could work. We use this for classification but not for...
Hey @shawnharmsen, thank you so much for contributing! I think this is exactly what we want! I'll merge it first thing in the morning after reviewing it. 👍 🔥 🚀
That is our demo on Gorilla Spotlight! Please join our waitlist here: https://forms.gle/T9P298Z4Pvok2v6MA and stay tuned! Thanks for your patience!
Thanks for your contribution!🚀 I have been able to process the file and get this new updated API. Could you review it and let me know if that is correct...
It is because the two questions you wrote is related to these two API calls and we just extract that from the entire JSON dataset. If you could come up...
1. Sorry that we are planning to go through the reviewing process to make sure everything looks ok on our end. It is also aimed to provide a high-quality dataset....
Yes we are working on it! Will do a comparison soon