Instructions to use mindchain/hopper-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use mindchain/hopper-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 mindchain/hopper-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf mindchain/hopper-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 mindchain/hopper-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf mindchain/hopper-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 mindchain/hopper-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf mindchain/hopper-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 mindchain/hopper-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf mindchain/hopper-GGUF:Q4_K_M

Use Docker

docker model run hf.co/mindchain/hopper-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use mindchain/hopper-GGUF with Ollama:
```
ollama run hf.co/mindchain/hopper-GGUF:Q4_K_M
```
Unsloth Desktop

How to use mindchain/hopper-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mindchain/hopper-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": "mindchain/hopper-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use mindchain/hopper-GGUF with Docker Model Runner:
```
docker model run hf.co/mindchain/hopper-GGUF:Q4_K_M
```

Lemonade

How to use mindchain/hopper-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull mindchain/hopper-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.hopper-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use mindchain/hopper-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 mindchain/hopper-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 mindchain/hopper-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mindchain/hopper-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mindchain/hopper-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 "mindchain/hopper-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"

hopper Q4_K_M (gemergt) — Studienvergleich, NICHT kommerziell nutzbar

⚠️ RESEARCH AND DEMO ONLY. Der Upstream-Adapter (HopitAI/hopper) wurde mit RACE-Passagen trainiert, deren Terms auf abgeleitete Daten übergehen — auch auf dieses Q4_Merge. Kein Produktions-/Kommerzeinsatz. Eine Version ohne die RACE-Passagen ist Upstream in Entwicklung; Austausch gemäß Swap-Plan geplant.


Upstream-Adapter	HopitAI/hopper @ `80262fe93c42df744578d7cd8c48726b3b668b99` (LoRA r16, alpha 32)
Basis	Qwen/Qwen3.5-4B @ `851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a` (= Card-Pin)
Merge	transformers+peft `merge_and_unload`, bf16, CPU (4,21 B, 32 Layer, qwen3_5_text)
Toolchain	llama.cpp Commit `9575389` + `--no-mtp`, bf16 → Q4_K_M (2,71 GB) — Master baut kaputte qwen3_5-GGUFs, Pin euren Commit!
Temperatures	per-kind (hopper.json): choice 0.7899 · noul 0.7531 · score 0.8997
JevBench	v1.4.2 Platz 2: 63.5 (I 48.0 · C 79.1) — beste Kalibrierung der offenen Modelle; ⚠️ FREMD-Methodik, nie Gate-Basis
Provenance	`provenance.json` mit beiden Pins + sha256

Warum wir ein eigenes Merge-Quant gebaut haben

Hopper ist nur als LoRA-Adapter veröffentlicht — lauffähig erst nach Merge in die Basis. Wir pinnen BEIDE Revisionen (Adapter + Basis), mergen reproduzierbar (Kernel) und quantisieren mit demselben Familien-Rezept wie decider.

Rolle im JEV-Stack

Studienvergleich (Kalibrierungs-Referenz gegen decider-4b v2, ECE 0,0374) und Gold-Set-Messung. Kein Router-/Produktions-Pfad vor dem Lizenz-Swap.

Downloads last month: 42

GGUF

Model size

4B params

Architecture

qwen35

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support