Instructions to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-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 F-Labs/Qwen3.5-0.8B-DynamicTune-Base-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 F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

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 F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama cli -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

Use Docker

docker model run hf.co/F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

Llama new
LM Studio
Jan
Ollama
How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF with Ollama:
```
ollama run hf.co/F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16
```
Unsloth Desktop

How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

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": "F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF with Docker Model Runner:
```
docker model run hf.co/F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16
```

Lemonade

How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

Run and chat with the model

lemonade run user.Qwen3.5-0.8B-DynamicTune-Base-GGUF-F16

List all available models

lemonade list

Hermes Agent

How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-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 F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

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 F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16

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 "F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF:F16" \
  --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"

Qwen3.5-0.8B-DynamicTune-Base-GGUF

Official GGUF release of F-Labs/Qwen3.5-0.8B-DynamicTune-Base, created via closed-form trajectory transfer from Qwen 3.5 4B without traditional backpropagation.

Files Available

Qwen3.5-0.8B-DynamicTune-Base-F16.gguf: Full FP16 precision GGUF (2.02 GB).

Benchmark Highlights (Zero-Shot Base)

Independent Cloud Benchmark: ARC-Challenge (Full 1,172 items on Datacenter NVIDIA L4 via official lm_eval 0.4.12):
- DynamicTune 0.8B (This Model): 42.15% acc_norm (±1.44%), 40.19% acc
- Stock Qwen3.5-0.8B Base: 37.50% acc_norm (+4.65% delta, Z=3.23σ significance)
- 25k SFT Distillation (Mythos-0.8B): 38.10% acc_norm
- Stock Qwen3.5-2B Base (Q8): 41.10% acc_norm
- TPN Bench Coordinator run: ce494664-d077-4ff1-8741-15cedabc434c
- Acknowledgments: Special thanks to TPN Bench (TaoFu Protocol) for providing the cloud NVIDIA L4 GPU infrastructure to independently verify this checkpoint.
- Note: Closed-form surgery solved locally in ~12 min on consumer hardware with 0 backprop; benchmark independently executed in the cloud on datacenter NVIDIA L4.
HellaSwag (400 tasks, Vulkan llama.cpp): 54.75% -> 55.25% (+0.50% gain)
Held-out NLL (30 tasks): 0.582 -> 0.519 (-10.8% loss drop)
Biomedicine NLL: -23.8%
Math & Logic NLL: -14.6%

Usage with llama.cpp

llama-cli -m Qwen3.5-0.8B-DynamicTune-Base-F16.gguf -p "Question: How does DNA replication initiate?\nAnswer:" -c 2048 -n 128

Downloads last month: 113

GGUF

Model size

1B params

Architecture

qwen35

Hardware compatibility

16-bit

Inference Providers NEW

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

Model tree for F-Labs/Qwen3.5-0.8B-DynamicTune-Base-GGUF

Base model

Qwen/Qwen3.5-0.8B-Base

Finetuned

F-Labs/Qwen3.5-0.8B-DynamicTune-Base

Quantized

(1)

this model