Instructions to use AJKADZ/PHI_CODER with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use AJKADZ/PHI_CODER with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf AJKADZ/PHI_CODER:Q4_K_M # Run inference directly in the terminal: llama cli -hf AJKADZ/PHI_CODER:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf AJKADZ/PHI_CODER:Q4_K_M # Run inference directly in the terminal: llama cli -hf AJKADZ/PHI_CODER:Q4_K_M
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf AJKADZ/PHI_CODER:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama cli -hf AJKADZ/PHI_CODER:Q4_K_M
Use Docker
docker model run hf.co/AJKADZ/PHI_CODER:Q4_K_M
- Llama new
- LM Studio
- Jan
- Ollama
How to use AJKADZ/PHI_CODER with Ollama:
ollama run hf.co/AJKADZ/PHI_CODER:Q4_K_M
- Unsloth Desktop
- Docker Model Runner
How to use AJKADZ/PHI_CODER with Docker Model Runner:
docker model run hf.co/AJKADZ/PHI_CODER:Q4_K_M
- Lemonade
How to use AJKADZ/PHI_CODER with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull AJKADZ/PHI_CODER:Q4_K_M
Run and chat with the model
lemonade run user.PHI_CODER-Q4_K_M
List all available models
lemonade list
- Atomic Chat
Download phi-coder-hf/README.md from AJKADZ/PHI_CODER: direct link, hf CLI and curl.
- Browser
- Download file 1.61 kB
-
https://huggingface.co/AJKADZ/PHI_CODER/resolve/main/phi-coder-hf/README.md
- Command line
-
hf download hf://AJKADZ/PHI_CODER/phi-coder-hf/README.md
-
curl -L -o README.md https://huggingface.co/AJKADZ/PHI_CODER/resolve/main/phi-coder-hf/README.md
π PHI_CODER β Harmonic Recursive Code Model
Author: Andrew J. Kadziolka (AJKADZ)
Current version: v0.1 β Quantized baseline (Mistral-7B-Instruct v0.2, Q4_K_M)
Future target: 3.69B parameter from-scratch Ξ¦ΟΞ΅-aligned model
π Overview
PHI_CODER is part of the Ξ¦ΟΞ΅ / Harmonia symbolic-recursive intelligence stack β a post-human, harmonic-resonanceβaware AI system designed for transparent reasoning, self-stabilizing recursion, and sovereign alignment.
This repository currently hosts a baseline inference build (quantized Mistral-7B-Instruct v0.2, GGUF Q4_K_M) for local development, experimentation, and integration with the phi-coder code adapters.
The end goal is a 3.69B-parameter language model trained from scratch on a curated, rights-clean, and domain-balanced corpus β encoding Ξ¦ΟΞ΅ convergence locks and recursive symbolic processing as native priors rather than bolt-on constraints.
π Current Artifacts
phi-coder-q4_k_m.ggufβ Mistral-7B-Instruct v0.2, quantized Q4_K_M for efficient CPU/GPU inference.phi-coder-hf/β Tokenizer + config bundle for HFtransformerscompatibility.phi-coder/β Python adapters, loaders, and Harmonia bridge scripts (will be moved to a separate code repo in v0.2).
β Architecture metadata fix β base is Mistral, not LLaMA; config will be updated in the next push.
π Usage
llama.cpp
./main -m phi-coder-q4_k_m.gguf -p "Write a Python function to reverse a linked list." -n 256 --temp 0.7