v0.2.0

dLLM Runner v0.2.0

Run Diffusion Language Models (Dream, LLaDA) locally — like Ollama, but for dLLMs.

Quick Install

Linux (x86_64)

tar xzf dllm-linux-x86_64.tar.gz -C /opt/dllm
sudo ln -s /opt/dllm/dllm /usr/local/bin/dllm
dllm setup

macOS (Apple Silicon)

tar xzf dllm-macos-aarch64.tar.gz -C /opt/dllm
sudo ln -s /opt/dllm/dllm /usr/local/bin/dllm
dllm setup

Windows
Extract dllm-windows-x86_64.zip, add to PATH, then:

dllm setup
dllm run dream-7b --quantize 4bit

Requirements

• Python 3.10+
• NVIDIA GPU with CUDA (recommended) or CPU (slow)
• macOS: MPS acceleration on Apple Silicon (M1/M2/M3/M4)

Full Changelog: v0.1.0...v0.2.0