ml/README.md

@ml - Machine Learning Packages

Workspace for ML infrastructure packages including GPU/RAM coordination, model loading, and memory management.

Packages

Resource Coordination

Package	Type	Purpose
vram-boss-py	Python	GPU/VRAM lease coordination
vram-boss-ts	TypeScript	VRAM coordination client
ram-boss-py	Python	RAM lease coordination + cache management
ram-boss-ts	TypeScript	RAM coordination client

Model Loading

Package	Type	Purpose
model-boss-py	Python	Unified model loading (HF, GGUF, Diffusers, etc.)
model-boss-ts	TypeScript	Model path resolution

CLI Tools

Package	Location	Purpose
bitch	@cli/bitch-cli	Unified CLI for RAM/VRAM management

Quick Start

CLI (Recommended)

# Install global CLI
npm install -g @lilith/bitch --registry=http://forge.nasty.sh/api/packages/lilith/npm/

# Install Python backends
pip install lilith-vram-boss lilith-ram-boss

# Use unified interface
bitch vram status      # GPU coordination
bitch ram analyze      # RAM management

Python Development

# Install packages
pip install lilith-vram-boss lilith-ram-boss lilith-model-boss

# Use in code
from lilith_vram_boss import GPUBoss
from lilith_ram_boss import RAMBoss
from lilith_model_boss import ManagedModelLoader

Documentation

• QUICK_START.md - Installation and common tasks
• ARCHITECTURE.md - Package architecture and usage patterns
• IMPLEMENTATION_SUMMARY.md - Implementation details

Architecture

bitch CLI (unified interface)
    ├── bitch vram <command>  →  vram-boss (Python)
    └── bitch ram <command>   →  ram-boss (Python)

model-boss (model loading)
    └── uses vram-boss internally for GPU coordination

Common Tasks

# Check resources
bitch vram status
bitch ram status

# Memory analysis
bitch ram analyze --processes --leaks

# Cache cleanup
bitch ram clear auto

# GPU coordination
bitch vram drain
bitch vram cleanup

License

MIT