gemma

Fine tune Gemma 3 on an object detection task

A private Claude-Code-style coding agent for Apple Silicon — run chat, code, and local model workflows on-device. MLX-native, Ollama/OpenAI API compatible, zero API keys.

Self-host a ChatGPT-style web interface for Ollama 🦙

Gemma models running 100% local on an 8 GB Jetson Orin Nano: voice assistants with RAG, multi-agent NPCs, live EN-JA translation, and a Gemma 4 voice + vision agent that decides when to look through the webcam. The demos from the Gemma 2 launch at Google Tokyo.

NAACL '24 (Best Demo Paper RunnerUp) / MlSys @ NeurIPS '23 - RedCoast: A Lightweight Tool to Automate Distributed Training and Inference

A lightweight implementation of the GEPA (Genetic-Pareto) prompt optimization method for large language models.

Train GEMMA on TPU/GPU! (Codebase for training Gemma-Ko Series)

An interactive educational game built for the Google Gemma 3n Impact Challenge.

Web interface for TranslateGemma, Google's open translation model

Pair Claude Desktop on Anthropic with Claude Code routed through Ollama. Visual walkthrough + copy-paste prompt that cuts your Claude Code bill ~90%.

Sparse-upcycle a dense Gemma-3-270m into a Mixture-of-Experts (gemma3moe) and serve it on llama.cpp (CPU). Upcycle, train, inspect routing, GGUF.

OSINTai v4: local-first OSINT crawling, evidence analysis, correlation, timelines, and reporting.

Composable Python framework for LLM agent workflows focused on simplicity and developer experience. We ❤️ OSS.

Enterprise-grade Knowledge Distillation framework for Persian Mental Health analysis. Distills reasoning capabilities from Llama-3-70b into a local Gemma-3-1b model using QLoRA and synthetic Chain-of-Thought datasets.

Instruction/chat prompts creation library for text generation LLMs. It supports local and Hugging Face models.

Natural language control for Python CLI tools using locally-trained SLMs (CPU inference)

One Mac process. Many models. Real speed. Multi-model LLM serving with prefix reuse, MTP acceleration, and OpenAI APIs — built for Apple Silicon, measured against mlx-lm and llama.cpp.

Local CLI for Google's TranslateGemma translation models with multi-platform support (MLX for Apple Silicon, PyTorch for CUDA/CPU).

Talk to your coding agents — press a hotkey, speak your prompt, and it lands in Claude Code, Codex, or Cursor. Fully local voice dictation (Parakeet + Gemma), nothing leaves your machine. Windows, macOS & Linux.

Practical guide to fine-tuning LLMs like Gemma, LLaMA, and Mistral on custom datasets using popular frameworks.