Tiemu
An evolutionary simulation program inspired by Honkai: Star Rail's Amphoreus. In this digital cosmos, AI entities evolve across millions of generations, searching for an answer—be it 42 or 33550336—while trying to avoid the ultimate of becoming an Irontomb? Maybe.
Install / Use
/learn @YezQiu/TiemuREADME
Tiemu Project
Quick Start
-
Clone the repository:
git clone https://github.com/YezQiu/Tiemu.git cd TiemuOr use whatever method to get the whole thing into your computer
-
Create and activate a virtual environment:
python3 -m venv .venv source .venv/bin/activate -
Install dependencies:
pip install -r requirements.txt -
Download the model file:
- Download
Qwen3-0.6B-Q8_0.gguffrom Hugging Face or the official source. - Place it in the
modelsfolder.
- Download
-
Run the project:
python main.pyNo lama_cpp:
python main.py --disable-llm
Notes
- The
modelsfolder and large model files are ignored by git. - If you encounter issues, check that all dependencies are installed and the model file is present.
- you should still be able to run the thing without the LLM model.
- you can run without LLM whell package, use arg:
--disable-llm. - This project uses
pyclogfor logging. See About the Logging System for a quick guide, and the Pyclog Project for full details.
Interactive Console and Archiving
This project has a powerful built-in interactive console that allows you to pause the simulation, observe, and even modify the world's parameters in real-time. Additionally, you can save and load the simulation progress at any time.
Interactive Console
During the simulation, you can press the p key at any time to pause the simulation and enter the interactive console. Here, you can observe and intervene in the evolution of Omphalos like a "La Lou ke" (辣卤客).
The console will display the prompt (翁法罗斯创世涡心) >, waiting for you to input commands.
List of Available Commands:
corcontinue: Resume the simulation.nornext: Execute the next generation (one frame), then pause again. This is very useful for observing evolution frame by frame.status: Display the macro status of the world, including the current generation, population size, ecological diversity, stagnation count, etc.p <entity_name|baie>: Print detailed information of the specified entity or the current "Baie" (白厄), including its Titan affinity and path tendency.-
Example:
p Neikos-0496
-
top [k]: Display the top k entities with the highest scores in the current world (default is 5).-
Example:
top 10
-
zeitgeist: View the current mainstream ideology and its weight as determined by the Citizens' Council.blueprint: View the current global evolutionary blueprint (base Titan affinities).set <parameter_name> <value>: Core control function. Dynamically modify a simulation parameter.-
Example:
set mutation_rate 0.5(Immediately adjusts the mutation rate to 0.5)
-
save <filename.json>: Save the current simulation state completely to a file.load <filename.json>: Load a simulation state from a file, completely overwriting the current world.help: Display help information for all available commands.
Save & Load
You can utilize the save/load functionality in two ways:
- During the simulation
After entering the interactive console by pressing the p key, you can use the following commands:
-
Save current progress:
save my_world_state.json -
Load previous progress:
load my_world_state.jsonNote: After loading, all current progress will be overwritten by the archive. You need to enter
cornto continue running the loaded world. -
Start from an archive
You can directly load an archive file via a command-line argument when launching the project. This is very convenient for continuing a previous simulation.
Thanks
- HeyCrab3
- AvalonRuFae
- MoonofBridge24
- Akanyi
Related Skills
node-connect
349.0kDiagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps
frontend-design
109.4kCreate distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
openai-whisper-api
349.0kTranscribe audio via OpenAI Audio Transcriptions API (Whisper).
qqbot-media
349.0kQQBot 富媒体收发能力。使用 <qqmedia> 标签,系统根据文件扩展名自动识别类型(图片/语音/视频/文件)。
