Open
Conversation
…es (10% replacement rate per generation).
…(N=dim(neural-network-parameters)). It's SLOW but maybe results are now better.
…timated gradient using ES with number of population/iterations N = dim(neural-network-parameters) as recommended by ChatGPT.
…oodness in a group of agents (plays also against other agents and not just itself).
…= (o-mu)*a) so batch norm parameters are linear and can be polyak averaged more easily.
… polyak averaging of Q and policy network parameters (BatchNorm is only enabled in non-recurrent reinforcement model RIFL2 and not RIFL4)
…e it doesn't seem to work only with RIFL2).
…ion parameters after gradient step even if solution doesn't improve.
…Y_SIZE number of most recent samples.
… to not to improve.
…rk a little better in practice hard problems.
… (FIXME: test that reinforcement learning still works with less initial iterations?)
…Cart Pole problem anymore or forgot the found solution quickly.
…e problem (doesn't seem to learn it???).
…uggy? + don't save/load etc). Reinforcement learning RIFL2 updates to support priority sampling (actions causing large changes preferred) of replay buffer. RIFL4 don't work for now.
…caused restarting to fail to continue where it was.
…k correctly with LayerNorm. (jacobian matrix layernorm calculation can be still buggy?).
…T WORK. So: DO NOT USE LAYER NORM.
…ision or something..
…ing of tau values.
…ee if there are bugs.
…cent (creating training and testing dataset error if first try didnt succeed).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Pull request