fattummy 0.1.0

A declarative, ultra-minimalist ML framework for zero-boilerplate hardware-agnostic inference and training.

FatTummy

A declarative, ultra-minimalist Python framework designed to collapse complex data processing, hardware detection (GPU/TPU), multi-engine inference (APIs + Local), fine-tuning, and custom architecture deployment into a beautiful, stateless 5-command interface.

Installation

pip install fattummy

Quick Start

Building and chatting with a custom 10B Mixture of Experts model:

import fattummy as ft

ft.build()
ft.modelbuild("10B")
ft.type(ft.MOOE)
ft.data("bigcode/the-stack-v2", "bigcode/starcoderdata")
ft.temp(0.7)
ft.chat()

The framework automatically detects if you are on a TPU or GPU, installs the correct PyTorch wheels natively, configures Hugging Face dependencies, and launches an interactive chat session.