Kata_web is an advanced CPU-based Go AI engine, utilizing the Custom Trained KW serie Models which uses larger LR for efficient performance and competitive level usage. This project is designed to provide free, cross-platform competitive AI and AI models for advanced Go competing while keeping the full source code private.
This project is licensed under the terms specified in the LICENSE file. By using, modifying, or distributing this software, you must:
- Comply with the license terms (review the LICENSE file).
- Give explicit credit to the original author.
- Seek permission for modifications/redistribution.
# -------------------------------------------------------------------
# Based on Kata_web (2025 Cheng Chang)
# Original source: https://github.com/changcheng967/Kata_web
# License: https://github.com/changcheng967/Kata_web/blob/main/LICENSE
# -------------------------------------------------------------------- AI-powered strategic insights – Leverages Kata_web self-trained custom competitive level b28c512nbt model for superior game analysis.
- Eigen backend for CPU support – Optimized for non-GPU environments.
- Cross-platform compatibility – Runs efficiently on almost every hardware.(Linux x64, linux ARM, Windows x64, Windows ARM)
- Legacy yet functional – Free for community use, though newer models exist.
git clone https://github.com/changcheng967/Kata_web.git
cd Kata_web
pip install -r requirements.txt
python main.py- Attribution: Always credit the original author (e.g., in code, docs, or UI).
- License Compliance: Refer to the LICENSE for legal terms.
- Modifications: Contact the author for derivative works.
Note: This is the public version of the code which is not the full version(minimal for usage) The private version will remain private unless we are told to open source the full code by Doulet Media.