Instructions to use Auroraventures/cipher-sft-merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Auroraventures/cipher-sft-merged with Transformers:

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

pipe = pipeline("text-generation", model="Auroraventures/cipher-sft-merged")
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, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("Auroraventures/cipher-sft-merged")
model = AutoModelForMultimodalLM.from_pretrained("Auroraventures/cipher-sft-merged")
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 Settings

vLLM

How to use Auroraventures/cipher-sft-merged with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Auroraventures/cipher-sft-merged

SGLang

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

Unsloth Studio

How to use Auroraventures/cipher-sft-merged 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 Auroraventures/cipher-sft-merged 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 Auroraventures/cipher-sft-merged to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Auroraventures/cipher-sft-merged to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="Auroraventures/cipher-sft-merged",
    max_seq_length=2048,
)

Docker Model Runner
How to use Auroraventures/cipher-sft-merged with Docker Model Runner:
```
docker model run hf.co/Auroraventures/cipher-sft-merged
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Cipher SFT (v1 base) 🦑

First taste of the Kraken. The creative-web code generator that started it all.

Cipher-SFT-Merged is the v1 SFT checkpoint of the Cipher series — a 31 B Gemma-4 fine-tune that writes complete, single-file HTML documents in the Awwwards idiom. This checkpoint provides the foundation that all later Cipher stages (SimPO, SFT 2.5, GRPO) are built on.

🧠 Base: unsloth/gemma-4-31b-it-unsloth-bnb-4bit
🔬 Fine-tune: Unsloth QLoRA SFT on early Awwwards-style code pairs
🎨 Specializes in: hero sections, scroll experiences, glassmorphism, WebGL, Three.js, GSAP
⚙️ Merged weights: BF16, ready for HF Inference Endpoints or llama.cpp via the GGUF sibling

🔄 Upgrade path: For the best Cipher experience, see cipher-sft25-real-merged — the v3 checkpoint trained on real scraped Awwwards code.

Quickstart

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

tok = AutoTokenizer.from_pretrained("Auroraventures/cipher-sft-merged")
model = AutoModelForCausalLM.from_pretrained(
    "Auroraventures/cipher-sft-merged",
    torch_dtype=torch.bfloat16, device_map="auto",
)

messages = [
    {"role": "system", "content": "You are Cipher, the Code Kraken. Emit complete single-file HTML."},
    {"role": "user", "content": "Build a portfolio landing with Lenis smooth scroll and GSAP reveals."},
]
prompt = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tok(prompt, return_tensors="pt").to(model.device)
print(tok.decode(model.generate(**inputs, max_new_tokens=4096, temperature=0.7)[0]))

Deploy as HF Inference Endpoint

Click Deploy → Inference Endpoints (dedicated) at the top of this page. Recommended: AWS us-east-1, Nvidia L4 ×1, min replicas 0. Then call via /v1/chat/completions (TGI OpenAI-compatible route).

Cipher Pipeline Map

Stage	Model	Status
SFT v1 (base)	cipher-sft-merged	✅ Apr 15 (this card)
SimPO anti-slop	cipher-simpo-merged	✅ Apr 15
SFT 2.5 synth	cipher-sft25-merged	⚠️ Retired — template collapse
SFT 2.5 real	cipher-sft25-real-merged	✅ Apr 18 — recommended
GRPO	planned	🔮 Q2 2026
KTO	planned	🔮 Q2 2026

License

CC-BY-NC-4.0. Base model governed by Gemma Terms of Use.

Built with 🦑 by Aurora Ventures. Part of the Kin runtime.

Downloads last month: 4

Safetensors

Model size

33B params

Tensor type

BF16

Model tree for Auroraventures/cipher-sft-merged

Finetunes

1 model

Quantizations