-
I would like to know the size of the buffer used in the SMAC scenario in the paper, as I reproduced the experiment with a buffer size of 100,000 on the 3m_good dataset and found that the performance o…
-
Thank you very much for your outstanding contributions in this field. I have some questions about installing related packages.
tensorstore 0.1.54 requires ml-dtypes>=0.3.1.
But tensorflow-intel …
-
你好，在训练结束后出现DataHandler: unable to parse websocket frame.
CloseHandler: 127.0.0.1:52886 disconnected
ResponseThread: No connection, dropping the response.
这个如何解决呢？
> @clb-Lenovo-Rescuer-15ISK:~/桌面/…
-
### What happened + What you expected to happen
There is a ~40% throughput drop between 6/9 and 6/10, commits 6f3de2af863cff95ff55a587003faf2b776fec65 and d0f975e00f32c5f632a39fbe8efb6df678f8b0a0. Th…
-
### What is the problem?
When training in a **multi-agent** environment using **multiple environment workers**, the memory of the **workers** increases constantly and is **not released after the poli…
-
Your q-learning baselines assume all agents have same obs and action dim, however in MPE_Single_Tag there are two different obs sizes.How can i apply qmix.py on MPE_Single_Tag?
-
你好，我有个问题想问一下您，在generate_episodeRNN中的隐层一直在变化，当生成一个episode，此时RNN中的隐层是最后一个step产生的隐层，那么在接下来的train中，在qmix.py的get_q_values中也使用到了这个隐层，但此时RNN中的隐层是最后一个step产生的隐层，这样是否合适？是否应该使用init_hidden后的隐层？
期待您的回复。
-
Hi!
Thanks a lot for this very interesting line of work. I am particularly curious about the zero-shot transfer learning performance shown in your paper.
However, I don't understand how to "appl…
-
Why `all_mixed_ood_qs.append(chosen_action_qs) # [B, T, Ra + 1]` in og_marl/tf2/systems/qmix_cql.py line 170.
I am a little confused about this, since "chosen_action_qs" is not a "ood_action", b…
-
官方结果显示finetuned-qmix的数据如图，但是我看论文中的配图在这几个场景中貌似数据有些差异，如5m-vs_6m，文中曲线貌似在0.75左右。这个胜率的统计是怎么得出的呢，是因为曲线图做了平滑处理吗，实际胜率直接取最大胜率？
![b2a1061b076315ece150bf7d031611a](https://github.com/tjuHaoXiaotian/pymarl3/asse…