Skip to content

ExaRL tasks - #241

Open
Ted Fujimoto (tcfuji) wants to merge 14 commits into
developfrom
develop_tf
Open

ExaRL tasks#241
Ted Fujimoto (tcfuji) wants to merge 14 commits into
developfrom
develop_tf

Conversation

@tcfuji

@tcfuji Ted Fujimoto (tcfuji) commented Nov 2, 2022

Copy link
Copy Markdown
  1. Made update_target method agent specific and not exposed to workflow. It's now used inside the train call.
  2. Made a common folder and separated out models and replay buffers. Agent vault only has agents.
  3. Moved epsilon to agents.
  4. Include PER beta.

This PR builds off of the develop_sept_22 branch.

Jodasue and others added 8 commits September 24, 2022 15:39
2. Fixed support for -A option (previously not being respected)
3. Ensured Tensorflow is not loaded when using Pytorch
4. Fixed log/results directory to increment instead of defaulting to
RUN001
5. Added convergence cutoffs to workflows
6. Fixed/simplified step counting in workflows
7. Added hyper-parameter tuning using sbatch and optuna
2. Added fix to sync with episode block leading to deadlock
3. Commented hyper-parameter tuning script
4. Fixed RUN000 to the next dir if logs exist
5. Adjusted output to match hyper-parameter script
…t in send_model method and model_count in sync_learning.py
2. Minor fix to bsuite_batch help message
2. Fixed print and end of run
3. Async convergence call had too many parameters
…to dqn.py. Need to resolve sample weight issue (line 392).
…ad of a beta parameter. Also mentioned other troubling areas in dqn.py using TODOs.
2. Updating analyze reward to match the logs without epsilon
3. Fixing plotille to ignore data before complete roll.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants