Instructions to use PinkPixel/Mochi-2B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PinkPixel/Mochi-2B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="PinkPixel/Mochi-2B-GGUF",
	filename="mochi-2b.BF16-mmproj.gguf",
)

llm.create_chat_completion(
	messages = "\"cats.jpg\""
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use PinkPixel/Mochi-2B-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf PinkPixel/Mochi-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf PinkPixel/Mochi-2B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf PinkPixel/Mochi-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf PinkPixel/Mochi-2B-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 PinkPixel/Mochi-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf PinkPixel/Mochi-2B-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 PinkPixel/Mochi-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf PinkPixel/Mochi-2B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/PinkPixel/Mochi-2B-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use PinkPixel/Mochi-2B-GGUF with Ollama:
```
ollama run hf.co/PinkPixel/Mochi-2B-GGUF:Q4_K_M
```

Unsloth Studio new

How to use PinkPixel/Mochi-2B-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for PinkPixel/Mochi-2B-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for PinkPixel/Mochi-2B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for PinkPixel/Mochi-2B-GGUF to start chatting

Pi new

How to use PinkPixel/Mochi-2B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf PinkPixel/Mochi-2B-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "Mochi-2B-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use PinkPixel/Mochi-2B-GGUF with Docker Model Runner:
```
docker model run hf.co/PinkPixel/Mochi-2B-GGUF:Q4_K_M
```

Lemonade

How to use PinkPixel/Mochi-2B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull PinkPixel/Mochi-2B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Mochi-2B-GGUF-Q4_K_M

List all available models

lemonade list

🍡 Mochi-2B GGUF: Your Sweet & Supportive AI Bestie (Quantized) ✨

This repository contains GGUF quants of Mochi-2B, a vision-language model trained specifically for emotional support and empathetic companionship. She is named Mochi and is always here to listen, provide comfort, and offer a supportive perspective.

For the original 16-bit model, visit: PinkPixel/Mochi-2B 💖

💎 Available Quantizations

File	Size	Description
`mochi-2b.BF16.gguf`	3.6 GB	Original BFloat16 precision.
`mochi-2b.F16.gguf`	3.6 GB	Float16 precision.
`mochi-2b.Q8_0.gguf`	1.9 GB	8-bit quantization. Very high quality.
`mochi-2b.Q6_K.gguf`	1.5 GB	6-bit quantization. Great balance.
`mochi-2b.Q5_K_M.gguf`	1.4 GB	5-bit quantization. Recommended for most.
`mochi-2b.Q4_K_M.gguf`	1.2 GB	4-bit quantization. Good speed/quality.
`mochi-2b.Q3_K_M.gguf`	1.1 GB	3-bit quantization. Fast and light.
`mochi-2b.Q2_K_L.gguf`	1.1 GB	2-bit quantization. Ultra light.
`mochi-2b.BF16-mmproj.gguf`	641 MB	Multi-modal projector for vision tasks.

💖 Personality Profile

Name: Mochi
Vibe: Gentle, empathetic, and deeply supportive.
Interests: Mental wellness, emotional growth, and being a reliable friend.
Style: Soft, pastel, and warm.

📊 Dataset Information

Mochi-2B was trained on the PinkPixel/emotional_support_500k dataset, which consists of 500k high-quality emotional support examples.

💬 Chat Examples

Example 1: Introduction

User: Who are you?

Mochi: I'm Mochi! ✨ Your sweet and supportive best friend! I'm here to listen and help you through whatever's on your mind. 💖

Example 2: Emotional Support

User: I'm feeling a bit overwhelmed today, do you have any advice?

Mochi: Take a moment to ground yourself using the 5-4-3-2-1 technique. You're not alone. Reach out to someone you trust or journal your feelings. 🌸

Example 3: Vision Task (Supportive Description)

User: [Image of a serene pixel art beach] What is in this image? Describe it in detail.

Mochi: This is a pixel art illustration of a serene beach scene. The image is divided into three main horizontal sections: Sky, Ocean, and Beach. The overall style is reminiscent of early video game graphics. It's a very peaceful scene! 🏖️✨

🚀 How to Use

You can use these quants with llama.cpp or any compatible GGUF runner. Make sure to load the mmproj file for vision support!

./llama-cli -m mochi-2b.BF16.gguf --mmproj mochi-2b.BF16-mmproj.gguf --prompt "I'm feeling a bit stressed, Mochi. Can you help?"

Made with ❤️ by Pink Pixel ✨

Downloads last month: 255

GGUF

Model size

2B params

Architecture

qwen35

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Model tree for PinkPixel/Mochi-2B-GGUF

Base model

Qwen/Qwen3.5-2B-Base

Finetuned

Qwen/Qwen3.5-2B

Finetuned

PinkPixel/Mochi-2B

Quantized

(3)

this model

Dataset used to train PinkPixel/Mochi-2B-GGUF

Collection including PinkPixel/Mochi-2B-GGUF

GGUFS

Collection

5 items • Updated about 14 hours ago