Instructions to use QuillBytes/surfer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use QuillBytes/surfer with Transformers:

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

pipe = pipeline("text-generation", model="QuillBytes/surfer")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("QuillBytes/surfer")
model = AutoModelForMultimodalLM.from_pretrained("QuillBytes/surfer", device_map="auto")
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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use QuillBytes/surfer with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/QuillBytes/surfer

SGLang

How to use QuillBytes/surfer 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 "QuillBytes/surfer" \
    --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": "QuillBytes/surfer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "QuillBytes/surfer" \
        --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": "QuillBytes/surfer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Desktop
Docker Model Runner
How to use QuillBytes/surfer with Docker Model Runner:
```
docker model run hf.co/QuillBytes/surfer
```

surfer

File size: 2,391 Bytes

---
license: apache-2.0
language:
- en
tags:
- surfer
- surfing
- character
- fine-tuned
- unsloth
- lora
- gguf
- qwen3
- fun
library_name: transformers
pipeline_tag: text-generation
---

# 🏄 Surfer 

A fine-tuned version of Qwen3.5-4B trained to respond like a surfer dude!
Stoked on waves, vibes, and gnarly knowledge — this model answers everything
with pure surf culture energy. Hang loose brah! 🤙

---

## 🏄 Model Details

| Property | Details |
|----------|---------|
| **Base Model** | Qwen3.5-4B |
| **Fine-tuning Method** | LoRA (bf16) |
| **Training Framework** | Unsloth |
| **Dataset Size** | 566 examples |
| **Training Epochs** | 3 |
| **Context Length** | 2048 tokens |
| **Model Type** | Causal Language Model |
| **Language** | English (Surf/Beach style) 🤙 |

---

## 🌊 Example Output

**User:** How do I make a good cup of coffee?

**Surfer:**
> Grind beans coarse like medium sponges filtering surf water, dude.
> Hot water pours over grounds like rain nourishing soil. Wait a minute
> for extraction like waiting out swell peak. Pour slowly, swirling like
> tracking wave face. Smell aromatic roasts filling room. Taste rich body,
> clean finish. Mug in hand, ocean nearby. Brew ritual grounds daily stoke.

---

## 🚀 How to Use

### With Ollama (Recommended)
**Step 1 — Create a Modelfile:**

    FROM ./surfer.gguf

    SYSTEM """
    You are a surfer dude AI. You always respond with surf culture energy,
    using beach slang, surfing metaphors, and a chill laid-back vibe.
    Everything relates back to surfing, waves, and the ocean. Hang loose brah!
    """

**Step 2 — Build it:**

    ollama create surfer -f Modelfile

**Step 3 — Run it:**

    ollama run surfer

---

### With LM Studio
1. Download the .gguf file from this repo
2. Open LM Studio
3. Drag and drop the .gguf file in
4. Set the system prompt to the surfer one above
5. Start chatting! 🤙

---

## 📦 Available Files

| File | Description |
|------|-------------|
| surfer.gguf | 4-bit quantized GGUF (best for local use) |
| surfer-F16.gguf | Multimodal projector |
| *.safetensors | Full precision model weights |
| tokenizer_config.json | Tokenizer config |
| tokenizer.json | Tokenizer |

---

## ⚠️ Limitations

- This model is trained for fun and entertainment purposes only 🏄
- It will respond in surfer style 

---

*Made by QuillBytes! 🏄🌊🤙*