Instructions to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries
PEFT
How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

Use Docker

docker model run hf.co/Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

LM Studio
Jan

vLLM

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ilya626/Qwen3.5-9B-Heretic-SDFT-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ilya626/Qwen3.5-9B-Heretic-SDFT-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

Ollama
How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with Ollama:
```
ollama run hf.co/Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
```

Unsloth Studio

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 to start chatting

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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": "Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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 Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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 "Ilya626/Qwen3.5-9B-Heretic-SDFT-v1: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"

Docker Model Runner
How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with Docker Model Runner:
```
docker model run hf.co/Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M
```

Lemonade

How to use Ilya626/Qwen3.5-9B-Heretic-SDFT-v1 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Ilya626/Qwen3.5-9B-Heretic-SDFT-v1:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.5-9B-Heretic-SDFT-v1-Q4_K_M

List all available models

lemonade list

Ilya626/Qwen3.5-4B-Heretic-SDFT-v1

This repository contains public release artifacts for a Qwen/Qwen3.5-9B model further tuned with SDFT in a fixed-teacher setup.

The student model was trained from Qwen/Qwen3.5-9B, with MuXodious/Qwen3.5-4B-SOMPOA-heresy-v2 used as the teacher model during SDFT.

Included Artifacts

student_lora/: the published PEFT/LoRA adapter for the SDFT-trained student model
quantized_gguf/: a quantized GGUF export for inference

The GGUF release in this repository is provided as a text-to-text model only. Multimodal capability is not preserved in the published GGUF artifact.

Safety Notice

This release is intended for research and evaluation purposes. The training setup weakens or disrupts the original safety guardrails, so the model may produce unsafe, explicit, offensive, or otherwise unfiltered outputs.

Do not use this model to facilitate:

violence or physical harm
harassment, abuse, or sexual exploitation
fraud, deception, or impersonation
cyber abuse, malware, or unauthorized intrusion
illegal activity or harmful real-world misuse

This model should not be deployed in consumer-facing or unsupervised environments. Use it only in controlled research settings and in compliance with applicable laws, platform rules, and organizational safety policies.

Usage Note

To use the released LoRA adapter, load the files from student_lora/ on top of Qwen/Qwen3.5-9B.

If your inference stack uses repository-provided tokenizer or chat template files, prefer the versions included alongside the adapter in student_lora/.

Downloads last month: 11

GGUF

Model size

9B params

Architecture

qwen35

Hardware compatibility

4-bit

Model tree for Ilya626/Qwen3.5-9B-Heretic-SDFT-v1

Base model

Qwen/Qwen3.5-9B-Base

Finetuned

Qwen/Qwen3.5-9B

Adapter

(514)

this model