-
In PPOv2 trainer.train(), # 4. compute rewards, when computing the rewards index, the sequence_lengths_p1 is used.
`actual_end = torch.where(sequence_lengths_p1 < rewards.size(1), sequence_lengths…
-
There seem to be several memory leaks as matches pile up. But one area that is causing some of it appears to be `_reward_buffer` in `EnvPlayer`. It holds references to `AbstractBattles` and doesn't ev…
-
In the 3.0 paths in random it should give (Emblem of valor: 40753) and in the heroic ones after the first dungeon it should give (Emblem of heroism: 40752) in which it is only giving gold as a reward.…
-
We aim to integrate global toast notifications into our project to indicate various actions. We will use the default Bootstrap toast in three modes: **success** (green), **error** (red), and **info** …
-
### Summary
Development of this grant consists of the following set of steps:
- Implementation of the server side of the application, which contains information regarding the projects that exist in …
-
After updating ComfyUI, the node fails, updating the node does not help, and it still cannot be used
-
Hello, thank you very much for open sourcing such a great project. I am running the code:
python experiments.py evaluate configs/IntersectionEnv/env.json using the command \
…
-
Reproducibility is a huge pain point to tackle and kudos for spearheading this initiative.
However implementing potentially sparsely documented papers with potentially implementation "details" that…
-
### Name
Super Easter Egg Altar
### Type of Misc Marker
other (please specify in "Any Other Info")
### Description
Location of the "super easter egg" altar added later to the game. Grants a rewar…
-
OH! So hitting the wrong side once isn't wrong until he commits by licking the Number of Lick times.
I had thought that an error wouldn't need to be repeated to be an error. A quick test lick is allo…