KataGo/Linux:修订间差异
跳到导航
跳到搜索
此页面具有访问限制。如果您看见此消息,则说明您没有权限访问此页面。
(→執行) |
(→執行) |
||
第79行: | 第79行: | ||
</syntaxhighlight> | </syntaxhighlight> | ||
透過VNC連進去 後 ,用 終端機(Terminal ,預設是xterm )後執行[[Lizzie]]: | |||
<syntaxhighlight lang="bash"> | <syntaxhighlight lang="bash"> |
2020年3月8日 (日) 09:32的版本
KataGo是一套电脑围棋软体。
基本环境
这边是用AWS上的Ubuntu 18.04,开g4dn.xlarge
测试的。
编译
先安装需要的套件:
sudo apt install -y build-essential cmake git libboost-filesystem-dev libgoogle-perftools-dev libzip-dev nvidia-compute-utils-435 nvidia-driver-435 nvidia-headless-435 nvidia-utils-435 ocl-icd-opencl-dev unzip zlib1g-dev
再来编译:
cd ~; git clone https://github.com/lightvector/KataGo.git; cd ~/KataGo/cpp; cmake . -DBUILD_MCTS=1 -DUSE_BACKEND=OPENCL -DUSE_TCMALLOC=1; make -j8
然后把KataGo的执行档丢到Lizzie目录下:
cd ~/KataGo/cpp; cp katago ~/Lizzie
设定
等下会使用Lizzie当作前端,所以这边会把KataGo的设定档丢到Lizzie的目录下:
cd ~/KataGo/cpp; cp configs/gtp_example.cfg ~/Lizzie/
下载最新的训练资料
这边会下载目前最强的20 blocks、30 blocks与40 blocks版本到Lizzie的目录下并且解开:
cd ~/Lizzie/; wget https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170-b40c256x2-s1349368064-d524332537.zip https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170-b30c320x2-s1287828224-d525929064.zip https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170e-b20c256x2-s2430231552-d525879064.zip; echo *.zip | xargs -n1 unzip
测试
先测试Nvidia显卡有没有生效(要记得先重开机让驱动程式生效):
nvidia-smi
接下来可以跑KataGo的测试:
cd ~/KataGo/; ./katago benchmark -model g170-b40c256x2-s1349368064-d524332537/model.bin.gz -config gtp_example.cfg
跑完测试后可以依照建议修改gtp_example.cfg
里的numSearchThreads
以加快速度。
快速安装
sudo apt update; sudo apt install -y build-essential cmake default-jdk git icewm libboost-filesystem-dev libgoogle-perftools-dev libzip-dev nvidia-compute-utils-435 nvidia-driver-435 nvidia-headless-435 nvidia-utils-435 ocl-icd-opencl-dev tightvncserver unzip zlib1g-dev; cd /tmp; wget https://github.com/featurecat/lizzie/releases/download/0.7.2/Lizzie.0.7.2.Mac-Linux.zip; cd ~; unzip /tmp/Lizzie.0.7.2.Mac-Linux.zip; git clone https://github.com/lightvector/KataGo.git; cd KataGo/cpp; cmake . -DBUILD_MCTS=1 -DUSE_BACKEND=OPENCL -DUSE_TCMALLOC=1; make -j8; cp katago ~/Lizzie; cp configs/gtp_example.cfg ~/Lizzie; cd ~/Lizzie; wget https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170-b40c256x2-s1349368064-d524332537.zip https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170-b30c320x2-s1287828224-d525929064.zip https://d3dndmfyhecmj0.cloudfront.net/g170/neuralnets/g170e-b20c256x2-s2430231552-d525879064.zip; echo *.zip | xargs -n1 unzip; sudo reboot
重开机后连进去再跑:
cd ~/Lizzie; ./katago benchmark -model g170-b40c256x2-s1349368064-d524332537/model.bin.gz -config gtp_example.cfg
执行
先把VNC跑起来:
tightvncserver -geometry 1680x1050 -depth 24
透过VNC连进去后,用终端机(Terminal,预设是xterm)后执行Lizzie:
cd ~/Lizzie; java -jar lizzie.jar
相关连结
外部连结
- 官方网站 (英文)