ThinkExponential AILab

###################################################################################
#                                                                                 #
#  _____ _     _       _    _____                                  _   _       _  #
# |_   _| |   (_)     | |  |  ___|                                | | (_)     | | #
#   | | | |__  _ _ __ | | _| |____  ___ __   ___  _ __   ___ _ __ | |_ _  __ _| | #
#   | | | '_ \| | '_ \| |/ /  __\ \/ / '_ \ / _ \| '_ \ / _ \ '_ \| __| |/ _` | | #
#   | | | | | | | | | |   <| |___>  <| |_) | (_) | | | |  __/ | | | |_| | (_| | | #
#   \_/ |_| |_|_|_| |_|_|\_\____/_/\_\ .__/ \___/|_| |_|\___|_| |_|\__|_|\__,_|_| #
#                                    | |                                          #
#                                    |_|                                          #
#                                                                                 #
#   ___  _____ _           _                                                      #
#  / _ \|_   _| |         | |                                                     #
# / /_\ \ | | | |     __ _| |__                                                   #
# |  _  | | | | |    / _` | '_ \                                                  #
# | | | |_| |_| |___| (_| | |_) |                                                 #
# \_| |_/\___/\_____/\__,_|_.__/                                                  #
#                                                                                 #
#   Local AI Model Deployments  ::  ThinkExponential                              #
#                                                                                 #
###################################################################################

Local AI model deployments using Docker. Supports the most popular open-source generative AI models and is compatible with Windows + WSL2 on both NVIDIA and AMD GPUs, and CPU.

Deployments

Deployment	Models	Port
ollama/	Llama 3.3, Mistral, Gemma 3, DeepSeek-R1, Qwen 2.5, Phi-4, CodeLlama, …	3000 (WebUI), 11434 (API)
stable-diffusion/	SD 3.5, SDXL, FLUX.1, SD 2.1	7860
moneyprinter/	AI Short Video Generation (MoneyPrinterTurbo)	8501 (WebUI), 8080 (API)

Quick start

1. Check your GPU

chmod +x scripts/check-gpu.sh
./scripts/check-gpu.sh

2. (First time) Install GPU drivers in WSL2

GPU	Script
NVIDIA	chmod +x scripts/setup-wsl-nvidia.sh && ./scripts/setup-wsl-nvidia.sh
AMD	chmod +x scripts/setup-wsl-amd.sh && ./scripts/setup-wsl-amd.sh

3. Start a deployment

LLMs with Ollama + Open WebUI

# NVIDIA
docker compose -f ollama/docker-compose.nvidia.yml up -d

# AMD
docker compose -f ollama/docker-compose.amd.yml up -d

# CPU only
docker compose -f ollama/docker-compose.cpu.yml up -d

Open http://localhost:3000 → pull a model (e.g. llama3.3) → start chatting.

Image generation with Stable Diffusion

# NVIDIA
docker compose -f stable-diffusion/docker-compose.nvidia.yml up -d

# AMD
docker compose -f stable-diffusion/docker-compose.amd.yml up -d

# CPU only (slow)
docker compose -f stable-diffusion/docker-compose.cpu.yml up -d

Open http://localhost:7860 to access the Stable Diffusion WebUI.

Short video generation with MoneyPrinterTurbo

# 1. Edit the config with your LLM + video API keys
nano moneyprinter/config.toml

# 2. Build and start (CPU only — no GPU required)
docker compose -f moneyprinter/docker-compose.cpu.yml up -d

Open http://localhost:8501 for the WebUI or http://localhost:8080/docs for the REST API.

Requirements

Windows host

• Windows 11 (21H2+) or Windows 10 (21H2+, build 19044+)
• Hardware virtualization enabled in BIOS/UEFI (Intel VT-x / AMD-V / SVM Mode)
• Docker Desktop 4.x+ with WSL2 backend enabled or Docker Engine installed inside WSL2
• Git for Windows to clone this repository, or clone it inside WSL2 with git clone
• NVIDIA: Driver ≥ 525 from nvidia.com/drivers
• AMD: Adrenalin 23.x+ from amd.com/support

Inside WSL2 (Ubuntu)

• Compose V2 (docker compose) — included with Docker Desktop or the Docker Engine Compose plugin
• NVIDIA: NVIDIA Container Toolkit (run scripts/setup-wsl-nvidia.sh)
• AMD: ROCm 6.x + WSL2 kernel 5.15+ (run scripts/setup-wsl-amd.sh)

AMD users: Docker Engine inside WSL2 is recommended over Docker Desktop — Docker Desktop does not correctly expose /dev/kfd to containers, which prevents ROCm GPU access.

See the docs/ folder for detailed guides, or each deployment's README.md for quick-reference instructions.

Documentation

Guide	Description
docs/environment-setup.md	Full WSL2 + Docker + GPU driver setup walkthrough
docs/ollama-usage.md	LLM usage with API examples for each model
docs/stable-diffusion-usage.md	Image generation with prompts, API examples, and parameter reference
docs/moneyprinter-usage.md	Short video generation workflow, prompt tips, and API examples

Repository structure

thinkexponential-ailab/
├── docs/
│   ├── README.md                      # Documentation index
│   ├── environment-setup.md           # WSL2, Docker, NVIDIA, AMD setup guide
│   ├── ollama-usage.md                # LLM usage and examples
│   ├── stable-diffusion-usage.md     # Image generation usage and examples
│   └── moneyprinter-usage.md         # Short video generation usage and examples
├── ollama/                            # LLM server + chat UI
│   ├── docker-compose.nvidia.yml
│   ├── docker-compose.amd.yml
│   ├── docker-compose.cpu.yml
│   └── README.md
├── stable-diffusion/                  # Image generation
│   ├── docker-compose.nvidia.yml
│   ├── docker-compose.amd.yml
│   ├── docker-compose.cpu.yml
│   └── README.md
├── moneyprinter/                      # AI short video generation
│   ├── Dockerfile
│   ├── docker-compose.cpu.yml
│   ├── config.toml
│   └── README.md
└── scripts/
    ├── check-gpu.sh                   # Detect GPU and recommend compose file
    ├── setup-wsl-nvidia.sh            # Install NVIDIA Container Toolkit in WSL2
    └── setup-wsl-amd.sh              # Install AMD ROCm in WSL2

License

This project is licensed under the Apache License 2.0.

Note: This license covers only the code, scripts, and documentation in this repository. AI models referenced or used by this project have their own licenses. Please review the license of any AI model before use.

Contributing

Contributions are welcome. Please read:

• CONTRIBUTING.md
• CODE_OF_CONDUCT.md
• SUPPORT.md
• SECURITY.md
• CHANGELOG.md

Credits

Core projects and upstream repositories

• ollama/ollama — local LLM runtime used by the Ollama deployments.
• open-webui/open-webui — chat UI used with Ollama.
• AUTOMATIC1111/stable-diffusion-webui — upstream Stable Diffusion WebUI project.
• openvinotoolkit/stable-diffusion-webui — OpenVINO-focused fork used by the CPU deployment path.
• andrewmcwatters/stablediffusion — repository used in the CPU Docker build path.
• Stability-AI/generative-models — upstream dependency used in the CPU Docker build path.
• crowsonkb/k-diffusion — upstream dependency used in the CPU Docker build path.
• sczhou/CodeFormer — upstream dependency used in the CPU Docker build path.
• salesforce/BLIP — upstream dependency used in the CPU Docker build path.
• harry0703/MoneyPrinterTurbo — AI short video generation tool used by the MoneyPrinterTurbo deployment.

Runtime images

• ollama/ollama — container image used for Ollama services.
• universonic/stable-diffusion-webui — container image used for NVIDIA and AMD Stable Diffusion deployments.

Toolchain and model sources

• Docker — container platform for all local deployments.
• NVIDIA Container Toolkit — NVIDIA GPU container runtime support.
• ROCm — AMD GPU compute stack used by AMD deployment paths.
• OpenVINO — CPU acceleration framework used by the CPU Stable Diffusion path.
• Hugging Face — model hosting source used in setup and examples.
• CivitAI — community model source referenced in usage guidance.

Credits are provided for attribution and transparency. Each third-party project, image, model, and service is governed by its own license and terms of use.