Chessboard AI

This program finds what pieces moved from an intial image of a chessboard and gives that move to Leela Chess Zero which is Open source neural network based chess engine (https://lczero.org/) with Maias Weights (https://github.com/CSSLab/maia-chess):

Visual representation of piece detection model

How to run?

Please change your config in .Example env in Cpu to match your Auth0 api settings and Lc0 config and rename '.Example env' to '.env'.

Broken right now

Nvidia GPU (Faster Runtime)

1. install docker on your system https://docs.docker.com/engine/install/
2. follow the guide at https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html and install for your system.
3. run docker build -t chessai . in directory Nvidia containing the Dockerfile
4. run docker run -p 9081:9081 --gpus all chessai

CPU (Slower Runtime)

1. install docker on your system https://docs.docker.com/engine/install/
2. run docker build -t chessai . in directory Cpu directory containing the Dockerfile
3. run docker run -p 9081:9081 --gpus all chessai

The API

Calling the API to grab the moves from the image example in python:

Set the strenght to the depth you want in this case it is set to 1.

import requests

url = "http://localhost:9081/upload" #set your server address

payload = {}
files=[
  ('file',('test1.png',open('$PATHTOFILE','rb'),'image/png')) # replace $PATHTOFILE to the file path of the image
]
headers = {
  'authorization': $YOUROAuth2ClientSecretGoesHere # replace this value with your auth0 key

response = requests.request("POST", url, headers=headers, data=payload, files=files)

print(response.text)

Reset the chessboard and setting the AI to play as white

This resets the entire board and starts a fresh game with option to let AI play as white

import requests

url = "http://localhost:9081/gamestat"  #set your server address

payload = {'reset': 'true',
'iswhite': 'true'} # change value iswhite to false if you are playing as white
files=[

]
headers = {
  'authorization': $YOURAuth0ClientSecretGoesHere # replace this value with your auth0 key
}

response = requests.request("POST", url, headers=headers, data=payload, files=files)

print(response.text)

Unrelated: how to export conda environment

conda env export --from-history > linux_environment.yml