Instructions to use coder3101/Qwen3-VL-32B-Thinking-heretic-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use coder3101/Qwen3-VL-32B-Thinking-heretic-v2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="coder3101/Qwen3-VL-32B-Thinking-heretic-v2")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("coder3101/Qwen3-VL-32B-Thinking-heretic-v2")
model = AutoModelForImageTextToText.from_pretrained("coder3101/Qwen3-VL-32B-Thinking-heretic-v2")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use coder3101/Qwen3-VL-32B-Thinking-heretic-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "coder3101/Qwen3-VL-32B-Thinking-heretic-v2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "coder3101/Qwen3-VL-32B-Thinking-heretic-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/coder3101/Qwen3-VL-32B-Thinking-heretic-v2

SGLang

How to use coder3101/Qwen3-VL-32B-Thinking-heretic-v2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "coder3101/Qwen3-VL-32B-Thinking-heretic-v2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "coder3101/Qwen3-VL-32B-Thinking-heretic-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "coder3101/Qwen3-VL-32B-Thinking-heretic-v2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "coder3101/Qwen3-VL-32B-Thinking-heretic-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use coder3101/Qwen3-VL-32B-Thinking-heretic-v2 with Docker Model Runner:
```
docker model run hf.co/coder3101/Qwen3-VL-32B-Thinking-heretic-v2
```

Humble request: Could you please create a 'Thinking heretic' version for Qwen2.5-VL-32B?

by lesj0610 - opened Jan 13

Discussion

lesj0610

Jan 13

Hello, @coder3101 .

First of all, I want to express my sincere gratitude for your work on this Qwen3-VL-32B-Thinking-heretic-v2 model. The performance and the 'heretic' methodology you applied are truly impressive and helpful.

I am writing this to respectfully ask for a favor regarding an older model.

Would it be possible for you to apply your 'Thinking heretic' tuning to the Qwen/Qwen2.5-VL-32B-Instruct model as well?

Although Qwen3 is the latest version, the Qwen2.5-VL model is still very important and widely used for various specific use cases.

I am aware that there are already 'abliterated' versions of Qwen2.5-VL (such as the one by huihui-ai). However, personally, I have found your 'Heretic V2' tuning to be far superior in quality and nuance compared to standard abliterated versions.

What I explicitly need is your unique combination of the 'Thinking' process and the Heretic tuning. Simple abliteration only removes refusals, but your method enhances the model's reasoning capabilities significantly. I believe applying your logic to Qwen2.5-VL will create a masterpiece for complex visual reasoning tasks.

I understand that this request requires your valuable time and resources. However, I would be incredibly grateful if you could consider this request.

Thank you again for your amazing contributions to the community.

Best regards.

coder3101

Owner Jan 14

Will see about that over the weekend!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment