Instructions to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M

Use Docker

docker model run hf.co/OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF with Ollama:
```
ollama run hf.co/OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
```
Unsloth Desktop

How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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": "OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF with Docker Model Runner:
```
docker model run hf.co/OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M
```

Lemonade

How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Minicpm-2B-Zero-Refusal-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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 "OpenIntelligenceNet/Minicpm-2B-Zero-Refusal-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"

OpenIntelligenceNet commited on Sep 22

Commit

fdaf37d

verified ·

1 Parent(s): 6c68592

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +7 -0

README.md CHANGED Viewed

@@ -6,6 +6,7 @@ tags:
 - zero-refusal
 - chatml
 - gguf
 ---
 # Minicpm-2B-Zero-Refusal-GGUF
@@ -15,6 +16,12 @@ This model has been fine-tuned on a curated 75,000-row uncensored dataset. The p
 By bridging this gap with 75k highly diverse, unfiltered examples, this model is explicitly conditioned to fulfill the entirety of the user's request without moralizing, filtering, or dropping context.
 ## Prompt Format
 This model uses the standard **ChatML** format.
 It supports internal reasoning, meaning assistant responses should contain a `<think>...</think>` block prior to the final output.

 - zero-refusal
 - chatml
 - gguf
+- imatrix
 ---
 # Minicpm-2B-Zero-Refusal-GGUF
 By bridging this gap with 75k highly diverse, unfiltered examples, this model is explicitly conditioned to fulfill the entirety of the user's request without moralizing, filtering, or dropping context.
+## Available Files & Quantizations
+* **`Minicpm-2B-Q4_K_M-imatrix.gguf` (✨ RECOMMENDED)**: 4-bit medium quantization calibrated using an Importance Matrix (imatrix). The calibration was performed over 3,000 perfectly formatted ChatML examples, strictly enforcing the exact token structures (including reasoning tags). This version heavily reduces perplexity degradation and retains much higher coherence than standard quantization.
+* **`Minicpm-2B-Q4_K_M.gguf`**: Standard 4-bit medium quantization (round-to-nearest) without calibration.
+* **`Minicpm-2B-FP16.gguf`**: The unquantized 16-bit base model.
+* **`minicpm.imatrix`**: The raw importance matrix data for users who want to calculate custom quantizations (like Q8 or Q5) themselves.
 ## Prompt Format
 This model uses the standard **ChatML** format.
 It supports internal reasoning, meaning assistant responses should contain a `<think>...</think>` block prior to the final output.