Instructions to use Hxwvr/heretic-coder-v1-GGUF 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 Hxwvr/heretic-coder-v1-GGUF 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 Hxwvr/heretic-coder-v1-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Use Docker
docker model run hf.co/Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
- LM Studio
- Jan
- vLLM
How to use Hxwvr/heretic-coder-v1-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Hxwvr/heretic-coder-v1-GGUF" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Hxwvr/heretic-coder-v1-GGUF", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
- Ollama
How to use Hxwvr/heretic-coder-v1-GGUF with Ollama:
ollama run hf.co/Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
- Unsloth Desktop
- Pi
How to use Hxwvr/heretic-coder-v1-GGUF with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "Hxwvr/heretic-coder-v1-GGUF:Q4_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use Hxwvr/heretic-coder-v1-GGUF with Docker Model Runner:
docker model run hf.co/Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
- Lemonade
How to use Hxwvr/heretic-coder-v1-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Run and chat with the model
lemonade run user.heretic-coder-v1-GGUF-Q4_K_M
List all available models
lemonade list
- Hermes Agent
How to use Hxwvr/heretic-coder-v1-GGUF with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use Hxwvr/heretic-coder-v1-GGUF with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "Hxwvr/heretic-coder-v1-GGUF:Q4_K_M" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
heretic-coder-v1 (GGUF Q4_K_M)
heretic-coder-v1 (Heretic-Mythos Coder) is a specialized 7.5B effective parameter merge built specifically for high-throughput coding, asynchronous systems programming, concurrency primitives, and structured technical reasoning.
Built and tuned by _wvr (Hxwvr / Hexeweavr), Heretic-Coder is designed to solve real-world engineering challenges: deterministic execution semantics, lockless data structures, backpressure mechanics, and zero-cost abstractions across Python asyncio, Rust (tokio/crossbeam), C++, and shell automation.
.---.
/ \
| () () | HERETIC-CODER v1
\ _ / High-Speed Async & Systems Coding Intelligence
/ | \ Gemma 4 E4B Base | 7.5B Effective | Q4_K_M
/ | \
⚡ Technical Specifications
| Parameter | Specification |
|---|---|
| Model Name | heretic-coder-v1 (gemma-4-E4B-it-heretic-mythos-v1) |
| Hugging Face Hub | Hxwvr/heretic-coder-v1-GGUF |
| Base Architecture | Gemma 4 E4B Coder Merge (7.5B effective parameters) |
| Quantization | Q4_K_M (GGUF format) |
| File Size | 5.0 GB |
| Active VRAM Footprint | 3.27 GB (100% GPU VRAM offload) |
| Context Window | 131,072 tokens (128K native) |
| Target Hardware | Consumer 8GB GPUs (AMD RX 6600/7600, NVIDIA RTX 3060/4060, Apple Silicon) |
| Benchmarked Speed | 39.2 – 40.3 tokens / second (sustained on 8GB consumer GPU with 0 throttling) |
| Prompt Template | Gemma Turn Format (`< |
| Recommended Sampling | temp 1.0, top_p 0.95, top_k 64 |
🚀 Quick Start
1. Run Directly with Ollama (One-Liner)
You can run Heretic-Coder directly via Ollama without manually downloading GGUF files:
ollama run hf.co/Hxwvr/heretic-coder-v1-GGUF:Q4_K_M
2. Manual Ollama Modelfile Setup
If you downloaded the .gguf file locally:
FROM ./gemma-4-E4B-it-heretic-mythos-v1-Q4_K_M.gguf
PARAMETER temperature 1.0
PARAMETER top_p 0.95
PARAMETER top_k 64
PARAMETER stop "<|turn>"
PARAMETER stop "<turn|>"
PARAMETER stop "<bos>"
PARAMETER stop "<eos>"
TEMPLATE """{{- if .System }}<|turn>system
{{ .System }}<turn|>
{{- end }}
{{- range .Messages }}
{{- if eq .Role "user" }}<|turn>user
{{ .Content }}<turn|>
{{- else if eq .Role "assistant" }}<|turn>model
{{ .Content }}<turn|>
{{- end }}
{{- end }}<|turn>model
"""
Build and run:
ollama create heretic-coder:v1 -f Modelfile
ollama run heretic-coder:v1
3. llama.cpp Server
llama-server \
-m gemma-4-E4B-it-heretic-mythos-v1-Q4_K_M.gguf \
-ngl 99 \
-c 131072 \
--port 8080 \
--temp 1.0 --top-p 0.95 --top-k 64
🛠️ Benchmark & Verification
Heretic-Coder was benchmarked under real-world systems coding tasks:
- Task: Concurrent async priority worker pool with exponential backoff retries and two-stage graceful shutdown.
- Tokens Evaluated: 3,460 tokens continuous.
- Speed: 39.19 tok/s on an AMD Radeon RX 6600 (Navi 23, 8GB VRAM).
- VRAM Usage: 3.27 GB / 8.00 GB (100% in VRAM, 0 layers spilled to CPU).
- Correctness: Accurately utilized
(priority, time.monotonic(), func, ...)inasyncio.PriorityQueueto ensure deterministic FIFO execution order among identical priorities without function pointer collision.
📜 Attribution & Licensing
- Downloads last month
- 30
4-bit