Instructions to use JoyFusionAI/gguf-components with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use JoyFusionAI/gguf-components 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 JoyFusionAI/gguf-components:Q8_0
# Run inference directly in the terminal:
llama cli -hf JoyFusionAI/gguf-components:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf JoyFusionAI/gguf-components:Q8_0
# Run inference directly in the terminal:
llama cli -hf JoyFusionAI/gguf-components:Q8_0

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 JoyFusionAI/gguf-components:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf JoyFusionAI/gguf-components:Q8_0

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 JoyFusionAI/gguf-components:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf JoyFusionAI/gguf-components:Q8_0

Use Docker

docker model run hf.co/JoyFusionAI/gguf-components:Q8_0

Llama new
LM Studio
Jan
Ollama
How to use JoyFusionAI/gguf-components with Ollama:
```
ollama run hf.co/JoyFusionAI/gguf-components:Q8_0
```
Unsloth Desktop

How to use JoyFusionAI/gguf-components with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf JoyFusionAI/gguf-components:Q8_0

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": "JoyFusionAI/gguf-components:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use JoyFusionAI/gguf-components with Docker Model Runner:
```
docker model run hf.co/JoyFusionAI/gguf-components:Q8_0
```

Lemonade

How to use JoyFusionAI/gguf-components with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull JoyFusionAI/gguf-components:Q8_0

Run and chat with the model

lemonade run user.gguf-components-Q8_0

List all available models

lemonade list

Hermes Agent

How to use JoyFusionAI/gguf-components with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf JoyFusionAI/gguf-components:Q8_0

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 JoyFusionAI/gguf-components:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use JoyFusionAI/gguf-components with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf JoyFusionAI/gguf-components:Q8_0

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 "JoyFusionAI/gguf-components:Q8_0" \
  --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"

gguf-components / README.md

Norton0924

Mirror GGUF weights for stable-diffusion.cpp on Android (Z-Image Turbo, FLUX.2 klein 4B, Anima Turbo 1.1)

d898e12 verified 29 days ago

preview code

Raw History Blame Contribute Delete

2.32 kB

---
license: apache-2.0
pipeline_tag: text-to-image
tags:
- gguf
- stable-diffusion.cpp
- text-encoder
- vae
- android
---
# gguf-components

Shared text encoders and VAEs, in the formats [stable-diffusion.cpp](https://github.com/leejet/stable-diffusion.cpp)
loads directly, used by JoyFusion's Android on-device inference. All files are byte-identical mirrors of their sources.

| File | Used by | Source | Size | SHA-256 |
|---|---|---|---|---|
| `Qwen3-4B-Instruct-2507-Q4_0.gguf` | Z-Image Turbo (text encoder) | [unsloth/Qwen3-4B-Instruct-2507-GGUF](https://huggingface.co/unsloth/Qwen3-4B-Instruct-2507-GGUF) | 2,375,773,280 | `e0ba675d86ab277c61701c6793659b2ae801d95e3be791464c321e6fbf613be2` |
| `Qwen3-4B-Instruct-2507-Q8_0.gguf` | Z-Image Turbo (text encoder) | same | 4,280,405,600 | `391c1e410fd9f4cf2de2b510273b56a84c19ce18f4fa3bfb3774031dac4ef068` |
| `Qwen3-4B-Q4_0.gguf` | FLUX.2 klein 4B (text encoder) | [unsloth/Qwen3-4B-GGUF](https://huggingface.co/unsloth/Qwen3-4B-GGUF) | 2,375,773,472 | `92f4a83cdfe691b216a949f915e47f35f0ed85b833aa03490c719160ac5cdf7a` |
| `Qwen3-4B-Q8_0.gguf` | FLUX.2 klein 4B (text encoder) | same | 4,280,405,792 | `eed555233267a33c7e8ee31682762cc7751b3f6d224039086e0e846f05fffa5d` |
| `Qwen3-0.6B-Base.Q8_0.gguf` | Anima (text encoder) | [mradermacher/Qwen3-0.6B-Base-GGUF](https://huggingface.co/mradermacher/Qwen3-0.6B-Base-GGUF) | 639,447,232 | `4b088f1793f6cba9f0c2f77ab835ef6734f205c2159168698c6e1a51b7df168a` |
| `ae.safetensors` | Z-Image (FLUX.1 VAE) | [Comfy-Org/z_image_turbo](https://huggingface.co/Comfy-Org/z_image_turbo) | 335,304,388 | `afc8e28272cd15db3919bacdb6918ce9c1ed22e96cb12c4d5ed0fba823529e38` |
| `flux2-vae.safetensors` | FLUX.2 klein (FLUX.2 VAE) | [Comfy-Org/vae-text-encorder-for-flux-klein-4b](https://huggingface.co/Comfy-Org/vae-text-encorder-for-flux-klein-4b) | 336,211,292 | `868fe7b343cc8f3a19dbcfcafbc3d5f888802be3f89bd81b65b3621a066ce8f3` |
| `qwen_image_vae.safetensors` | Anima (Qwen-Image VAE) | [circlestone-labs/Anima](https://huggingface.co/circlestone-labs/Anima) | 253,806,246 | `a70580f0213e67967ee9c95f05bb400e8fb08307e017a924bf3441223e023d1f` |

## License

Qwen3 (Alibaba), the FLUX.1 / FLUX.2 VAEs (Black Forest Labs) and the Qwen-Image VAE (Alibaba) are all released under the
Apache License 2.0 (see `LICENSE.md`).

	---
	license: apache-2.0
	pipeline_tag: text-to-image
	tags:
	- gguf
	- stable-diffusion.cpp
	- text-encoder
	- vae
	- android
	---

	# gguf-components

	Shared text encoders and VAEs, in the formats [stable-diffusion.cpp](https://github.com/leejet/stable-diffusion.cpp)
	loads directly, used by JoyFusion's Android on-device inference. All files are byte-identical mirrors of their sources.

	\| File \| Used by \| Source \| Size \| SHA-256 \|
	\|---\|---\|---\|---\|---\|
	\| `Qwen3-4B-Instruct-2507-Q4_0.gguf` \| Z-Image Turbo (text encoder) \| [unsloth/Qwen3-4B-Instruct-2507-GGUF](https://huggingface.co/unsloth/Qwen3-4B-Instruct-2507-GGUF) \| 2,375,773,280 \| `e0ba675d86ab277c61701c6793659b2ae801d95e3be791464c321e6fbf613be2` \|
	\| `Qwen3-4B-Instruct-2507-Q8_0.gguf` \| Z-Image Turbo (text encoder) \| same \| 4,280,405,600 \| `391c1e410fd9f4cf2de2b510273b56a84c19ce18f4fa3bfb3774031dac4ef068` \|
	\| `Qwen3-4B-Q4_0.gguf` \| FLUX.2 klein 4B (text encoder) \| [unsloth/Qwen3-4B-GGUF](https://huggingface.co/unsloth/Qwen3-4B-GGUF) \| 2,375,773,472 \| `92f4a83cdfe691b216a949f915e47f35f0ed85b833aa03490c719160ac5cdf7a` \|
	\| `Qwen3-4B-Q8_0.gguf` \| FLUX.2 klein 4B (text encoder) \| same \| 4,280,405,792 \| `eed555233267a33c7e8ee31682762cc7751b3f6d224039086e0e846f05fffa5d` \|
	\| `Qwen3-0.6B-Base.Q8_0.gguf` \| Anima (text encoder) \| [mradermacher/Qwen3-0.6B-Base-GGUF](https://huggingface.co/mradermacher/Qwen3-0.6B-Base-GGUF) \| 639,447,232 \| `4b088f1793f6cba9f0c2f77ab835ef6734f205c2159168698c6e1a51b7df168a` \|
	\| `ae.safetensors` \| Z-Image (FLUX.1 VAE) \| [Comfy-Org/z_image_turbo](https://huggingface.co/Comfy-Org/z_image_turbo) \| 335,304,388 \| `afc8e28272cd15db3919bacdb6918ce9c1ed22e96cb12c4d5ed0fba823529e38` \|
	\| `flux2-vae.safetensors` \| FLUX.2 klein (FLUX.2 VAE) \| [Comfy-Org/vae-text-encorder-for-flux-klein-4b](https://huggingface.co/Comfy-Org/vae-text-encorder-for-flux-klein-4b) \| 336,211,292 \| `868fe7b343cc8f3a19dbcfcafbc3d5f888802be3f89bd81b65b3621a066ce8f3` \|
	\| `qwen_image_vae.safetensors` \| Anima (Qwen-Image VAE) \| [circlestone-labs/Anima](https://huggingface.co/circlestone-labs/Anima) \| 253,806,246 \| `a70580f0213e67967ee9c95f05bb400e8fb08307e017a924bf3441223e023d1f` \|

	## License

	Qwen3 (Alibaba), the FLUX.1 / FLUX.2 VAEs (Black Forest Labs) and the Qwen-Image VAE (Alibaba) are all released under the
	Apache License 2.0 (see `LICENSE.md`).