Search Unity

Search Results

Search Again
Page 1 of 2
  1. unity_-DoCqyPS6-iU3A
    Post

    Resolved What PPO approach is used an more

    I would start simple. In this case, I would - limit the episode length, but give 1/"maximum episode length" as penalty every step (to encourage...
    Post by: unity_-DoCqyPS6-iU3A, Oct 29, 2021 in forum: ML-Agents
  2. unity_-DoCqyPS6-iU3A
    Post

    Question Lag spikes every 3-4 seconds in Unity play mode while training

    My answer is kind of late, sorry. Think about it this way: your observations will always be collected and then processed all at once. So it will...
    Post by: unity_-DoCqyPS6-iU3A, Sep 10, 2021 in forum: ML-Agents
  3. unity_-DoCqyPS6-iU3A
    Post

    Question Unable to reach perfection in extraordinarily simple environment

    I saw a video on YouTube on old version of mlagents where an episode length of 1 caused problems. Can you try making your episode at least 2 steps...
    Post by: unity_-DoCqyPS6-iU3A, Sep 10, 2021 in forum: ML-Agents
  4. unity_-DoCqyPS6-iU3A
    Post

    Question Lag spikes every 3-4 seconds in Unity play mode while training

    mlagents is probably processing all of the collected observations and training the neural net. During that time it can't provide Unity with new...
    Post by: unity_-DoCqyPS6-iU3A, Aug 27, 2021 in forum: ML-Agents
  5. unity_-DoCqyPS6-iU3A
    Post

    Summaries cause pauses in training with multiple agents

    Oh, SAC? I have no intuition about that algorithm, and don't have a clue what the problem could be, sorry.
    Post by: unity_-DoCqyPS6-iU3A, May 3, 2021 in forum: ML-Agents
  6. unity_-DoCqyPS6-iU3A
    Post

    Summaries cause pauses in training with multiple agents

    Can you please show us your config-file? Also, how many (combined) agents do you have across those 8 environments?
    Post by: unity_-DoCqyPS6-iU3A, Apr 26, 2021 in forum: ML-Agents
  7. unity_-DoCqyPS6-iU3A
    Post

    Question How many Unity Instances to run to optimize training speeds

    I just ran some benchmarks on my environment a few days ago. I reduced the max_steps in config.yaml to a small number so the training will...
    Post by: unity_-DoCqyPS6-iU3A, Apr 21, 2021 in forum: ML-Agents
  8. unity_-DoCqyPS6-iU3A
    Post

    Multiple Environments, Time Horizon, Batch-Size and On-Policy-Training

    Hello ruoping, thank you for the answers. Yeah, I had a few misconceptions about mlagents when I wrote that post, but I think I understand now....
    Post by: unity_-DoCqyPS6-iU3A, Apr 15, 2021 in forum: ML-Agents
  9. unity_-DoCqyPS6-iU3A
    Post

    Resolved Buffer.Shuffle - Performance-Gains for "sequence-length" of 1?

    Nvm, I'm not on the newest release and there have already been changes to this method in https://github.com/Unity-Technologies/ml-agents/pull/5192/
    Post by: unity_-DoCqyPS6-iU3A, Apr 12, 2021 in forum: ML-Agents
  10. unity_-DoCqyPS6-iU3A
    Thread

    Resolved Buffer.Shuffle - Performance-Gains for "sequence-length" of 1?

    Hello everyone, I was looking at the timers.json-file from my runs, specifically at the "_update_policy" section. "_update_policy" only has 1...
    Thread by: unity_-DoCqyPS6-iU3A, Apr 12, 2021, 1 replies, in forum: ML-Agents
  11. unity_-DoCqyPS6-iU3A
    Post

    Multiple Environments, Time Horizon, Batch-Size and On-Policy-Training

    I skimmed over the examples and "WallJump" has the following default-settings:...
    Post by: unity_-DoCqyPS6-iU3A, Apr 5, 2021 in forum: ML-Agents
  12. unity_-DoCqyPS6-iU3A
    Thread

    Multiple Environments, Time Horizon, Batch-Size and On-Policy-Training

    Hello everyone, I could use some insights into how ml-agents deals with trajectories that include observations from a previous policy (in my...
    Thread by: unity_-DoCqyPS6-iU3A, Apr 5, 2021, 4 replies, in forum: ML-Agents
  13. unity_-DoCqyPS6-iU3A
    Post

    What are the new "histograms" in tensorboard showing?

    Nevermind, it's probably the (average) number of agents that reached a certain reward at the end of the episode. No idea why I didn't think of...
    Post by: unity_-DoCqyPS6-iU3A, Feb 28, 2021 in forum: ML-Agents
  14. unity_-DoCqyPS6-iU3A
    Thread

    What are the new "histograms" in tensorboard showing?

    Release 13 added the "histogram" graph to tensorboard. But I can't figure out how the graph is connected to the generated rewards. Take this...
    Thread by: unity_-DoCqyPS6-iU3A, Feb 28, 2021, 2 replies, in forum: ML-Agents
  15. unity_-DoCqyPS6-iU3A
    Post

    Failed to build H5PY

    Had a problem with h5py yesterday as well. Solved it by using python 3.7.
    Post by: unity_-DoCqyPS6-iU3A, Feb 17, 2021 in forum: ML-Agents
  16. unity_-DoCqyPS6-iU3A
    Post

    (long term) Feature Request: Inference/Performance-Evaluation steps during training

    Yes! I only discovered it recently, but I find it *very* helpful.
    Post by: unity_-DoCqyPS6-iU3A, Sep 17, 2020 in forum: ML-Agents
  17. unity_-DoCqyPS6-iU3A
    Thread

    (long term) Feature Request: Inference/Performance-Evaluation steps during training

    Hello everyone, I've been tweaking my environment and playing with hyperparameters for quite some time, and finally reached a point where I'm...
    Thread by: unity_-DoCqyPS6-iU3A, Sep 17, 2020, 2 replies, in forum: ML-Agents
  18. unity_-DoCqyPS6-iU3A
    Thread

    Learning Rate, Epsilon and Policy Loss

    Hello everyone, this is a question to those users, who've managed to build up intuition about reading tensorboard-graphs. Or maybe those that...
    Thread by: unity_-DoCqyPS6-iU3A, Aug 8, 2020, 1 replies, in forum: ML-Agents
  19. unity_-DoCqyPS6-iU3A
    Thread

    Use C# files for debugging instead of dll?

    Hello everyone, is there an easy way to step through the source-code of ML-agents while debugging? Let's say I want to see the code of Agent.cs....
    Thread by: unity_-DoCqyPS6-iU3A, Aug 2, 2020, 1 replies, in forum: ML-Agents
  20. unity_-DoCqyPS6-iU3A
    Post

    Self-play and multi-agent reinforcement learning

    There is a bunch of links here...
    Post by: unity_-DoCqyPS6-iU3A, Jul 14, 2020 in forum: ML-Agents
Showing results 1 to 20 of 26 Next >
Search Again
Page 1 of 2