-
#91 を参照 (やっつけ試作あり).
この試作は, lz-analyze に以下の細工をしたもの.
* 探索が進むたびに, 探索した系列を「最善応手系列」と詐称して出力
* 適当に sleep をはさみながら途中経過を出力することで, 一手ずつ打っているようにアニメーション
GUI 側の対応が不要なのが利点 (Lizzie, LizzieYzy, LizGoban の「サブ碁…
-
Would you mind sharing your environment details so we can reproduce the result easily?
-
TamaGo を少し改造して MCTS の可視化を試しています. 現状のやっつけ実装からいくらか体裁を整えたら, マージの可能性はございますでしょうか?
教育やデモには良さそうに思われますが, プロジェクトの趣旨に合わなければご遠慮なく却下ください. 概念図ではない「現物」は見たことがなかったので, 個人的にはおもしろいです.
## アニメーション
https://github.…
-
Trying to debug larger width environments (7 currently).
Things to try:
1. Different metric (Average Q-value from 2015 paper https://arxiv.org/pdf/1312.5602.pdf).
```
5.1 Training and Sta…
-
## Motivation
It would be great to have an MCTS and Alphazero implementation, including other model-based RL for benchmarking and comparison.
## Solution
I can write a loss function of this po…
-
(rot) wangks@3dimage-31:/data/wangks/reflection-on-trees/reflection-on-trees$ sh blocksworld_rot.sh prompts/bw/pool_prompt_rot.json
Traceback (most recent call last):
File "/data/wangks/reflection…
-
Greeting, Dr!
I notice that from the input instance you share, we may have more than 1 depot. That means, we'll confront situation which we need to choose which depot to insert into the route (e.g.…
-
Hi, regarding the pseudocode for MCTS at line 96, it says "probability[i] = Math.exp(globalPolicy[currentStop][possible_successors[i]])".
May I know what does the currentStop stands for? Because my c…
-
I enabled MCTS ai, got this stack trace:
1) start game
2) choose yourself to go first
3) play a land
4) press spacebar
5) while AI is thinking press F4
Game exception occurred: java.lang.NullP…
-
@cooijmanstim
Hi, I am python programmer who started a [alphaGo Zero replication project](https://github.com/yhyu13/AlphaGOZero-python-tensorflow/tree/py2.7), I would like to practice a similar de…