Instructions to use daniloreddy/Qwen3.5-9B_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 daniloreddy/Qwen3.5-9B_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 daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf daniloreddy/Qwen3.5-9B_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 daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Use Docker

docker model run hf.co/daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use daniloreddy/Qwen3.5-9B_GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "daniloreddy/Qwen3.5-9B_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": "daniloreddy/Qwen3.5-9B_GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Ollama
How to use daniloreddy/Qwen3.5-9B_GGUF with Ollama:
```
ollama run hf.co/daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M
```
Unsloth Desktop

How to use daniloreddy/Qwen3.5-9B_GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf daniloreddy/Qwen3.5-9B_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": "daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use daniloreddy/Qwen3.5-9B_GGUF with Docker Model Runner:
```
docker model run hf.co/daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M
```

Lemonade

How to use daniloreddy/Qwen3.5-9B_GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.5-9B_GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use daniloreddy/Qwen3.5-9B_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 daniloreddy/Qwen3.5-9B_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 daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use daniloreddy/Qwen3.5-9B_GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf daniloreddy/Qwen3.5-9B_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 "daniloreddy/Qwen3.5-9B_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"

daniloreddy commited on Mar 14

Commit

e826d05

verified ·

1 Parent(s): 1f8a2d3

Update README.md

Browse files

Files changed (1) hide show

README.md +39 -25

README.md CHANGED Viewed

@@ -6,42 +6,56 @@ tags:
 - gguf
 - quantized
 - text-generation
-- qwen
-- instruct
-- chat
 ---
-# Qwen3.5-9B GGUF - High-Quality Quantizations
-This repository provides **GGUF** quantized versions of the [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) model, optimized for local execution using `llama.cpp`.
-> [!IMPORTANT]
-> **NOT A BASE MODEL:** These files are based on the standard **Qwen3.5-9B** release, which is the **Instruction Tuned (Chat)** version. It is specifically designed to follow instructions, engage in dialogue, and act as a conversational assistant.
-## Available Quantization Formats
 | File | Method | Bit | Description |
 | :--- | :--- | :--- | :--- |
-| **Q4_K_M.gguf** | Q4_K_M | 4-bit | **Recommended.** Best balance between speed and intelligence. |
-| **Q5_K_M.gguf** | Q5_K_M | 5-bit | High precision. Best for logical reasoning and structured output. |
-| **Q5_K_S.gguf** | Q5_K_S | 5-bit | High precision, slightly more efficient than Q5_K_M. |
-| **Q4_K_S.gguf** | Q4_K_S | 4-bit | Fast/Light. Optimized for maximum performance on CPU. |
-| **Q8_0.gguf** | Q8_0 | 8-bit | Near-lossless. Maximum fidelity to the original weights. |
-| **mmproj-f16.gguf** | F16 | 16-bit | **Vision Projector.** Required for image/video tasks. Keep in F16 for best accuracy. |
-## Quantization Details
-All models were quantized locally using the latest version of `llama.cpp`. K-Quants were used to ensure the best trade-off between perplexity and resource consumption on CPU-based systems.
-- **Quantization Date:** March 8, 2026
-- **Tooling:** `llama-quantize` (llama.cpp)
-- **Source Precision:** FP16 (Official weights)
-## Usage Instructions
-### 1. Multimodal Support (Vision/Image Analysis)
-To allow the model to "see" images, you need to load the Vision Projector (`mmproj`). This component translates visual data into a format the LLM can process.
-### 2. Running with llama-server (Recommended for GUI/API)
-To launch a local server with  vision support:
 ```bash
-./llama-server -m Qwen3.5-9B_Q4_K_M.gguf --mmproj mmproj-model-f16.gguf -n 512

 - gguf
 - quantized
 - text-generation
+- lightweight
+- lmstudio
+- jan
+- cobalt
+- text-generation-webui
 ---
+# Qwen3.5-9B - GGUF High-Quality Quantizations
+This repository provides **GGUF** quantized versions of the [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) model, optimized for local execution using `llama.cpp` and compatible ecosystems.
+## 📌 Version Notes
+All quantizations were generated from the official **FP16** weights.
+- **Target:** Efficient execution on consumer hardware, mobile/edge devices, and systems with limited memory.
+- **Performance:** The output quality (reasoning, coherence, and accuracy) is strictly dependent on the base model's parameter scale (9B).
+## 📊 Quantization Table
 | File | Method | Bit | Description |
 | :--- | :--- | :--- | :--- |
+| **fp16.gguf** | FP16 | 16-bit | **Original Weights.** No quantization applied. Maximum fidelity. |
+| **Q8_0.gguf** | Q8_0 | 8-bit | **Near-lossless.** Practically identical to the original model with lower memory footprint. |
+| **Q5_K_M.gguf** | Q5_K_M | 5-bit | **High Precision.** Minimizes quantization error for critical tasks. |
+| **Q4_K_M.gguf** | Q4_K_M | 4-bit | **Recommended.** Best balance between speed and performance. |
+| **Q4_K_S.gguf** | Q4_K_S | 4-bit | **Fast/Small.** Optimized for maximum throughput and low RAM usage. |
+## 🛠️ Technical Details
+- **Quantization Date:** 2026-03-07
+- **Tool used:** `llama-quantize` (llama.cpp)
+- **Method:** K-Quantization (optimized for AVX2/AVX-512 and modern GPU architectures).
+## 🚀 How to Use
+# Start a local OpenAI-compatible server with a web UI:
+### llama.cpp (CLI) using model from HuggingFace
+```bash
+./llama-cli -hf daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M -p "User: Hello! Assistant:" -n 512 --temp 0.7
+```
+### llama.cpp (CLI) using downloaded model
+```bash
+./llama-cli -m path/to/Qwen3.5-9B_Q4_K_M.gguf -p "User: Hello! Assistant:" -n 512 --temp 0.7
+```
+### llama.cpp (SERVER) using model from HuggingFace
+```bash
+./llama-server -hf daniloreddy/Qwen3.5-9B_GGUF:Q4_K_M --port 8080 -c 4096
+```
+### llama.cpp (SERVER) using downloaded model
 ```bash
+./llama-server -m /path/to/Qwen3.5-9B_Q4_K_M.gguf --port 8080 -c 4096
+```