Instructions to use decompute/Nebula-S-v1-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use decompute/Nebula-S-v1-4bit with Transformers:

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

pipe = pipeline("text-generation", model="decompute/Nebula-S-v1-4bit")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("decompute/Nebula-S-v1-4bit")
model = AutoModelForCausalLM.from_pretrained("decompute/Nebula-S-v1-4bit", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use decompute/Nebula-S-v1-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "decompute/Nebula-S-v1-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "decompute/Nebula-S-v1-4bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/decompute/Nebula-S-v1-4bit

SGLang

How to use decompute/Nebula-S-v1-4bit 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 "decompute/Nebula-S-v1-4bit" \
    --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": "decompute/Nebula-S-v1-4bit",
		"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 "decompute/Nebula-S-v1-4bit" \
        --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": "decompute/Nebula-S-v1-4bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use decompute/Nebula-S-v1-4bit with Docker Model Runner:
```
docker model run hf.co/decompute/Nebula-S-v1-4bit
```

Nebula-S-v1-4bit

4-bit quantized version of Nebula-S-v1.

Nebula-S-v1 is a reasoning-enhanced language model using the SVMS (Structured-Vector Multi-Stream) architecture.

What's different from Nebula-S-v1?

	Nebula-S-v1	Nebula-S-v1-4bit
Backbone precision	bf16	4-bit (nf4)
Adapter precision	bf16	bf16
Backbone size	~8 GB	~2 GB
Total size	~9 GB	~3 GB
VRAM needed	~18 GB	~6 GB
Requires	CUDA / MPS / CPU	CUDA only (bitsandbytes)

Quick Start

pip install torch transformers>=4.51.0 bitsandbytes accelerate huggingface-hub

Option 1: Using huggingface_hub

from huggingface_hub import snapshot_download
import sys

snapshot_download("punitdecomp/Nebula-S-v1-4bit", local_dir="./Nebula-S-v1-4bit")
sys.path.insert(0, "./Nebula-S-v1-4bit")
from nebula_s import load_nebula_s

model, tokenizer = load_nebula_s("./Nebula-S-v1-4bit", device="cuda")

Option 2: Using git clone

git lfs install
git clone https://huggingface.co/punitdecomp/Nebula-S-v1-4bit

import sys
sys.path.insert(0, "./Nebula-S-v1-4bit")
from nebula_s import load_nebula_s

model, tokenizer = load_nebula_s("./Nebula-S-v1-4bit", device="cuda")

Generate a response

messages = [{"role": "user", "content": "Solve step by step: what is 17 * 23?"}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to("cuda")
response = model.generate(
    inputs["input_ids"], inputs["attention_mask"],
    tokenizer, max_new_tokens=2048, temperature=0.7
)
print(response)

License

Apache 2.0. Backbone derived from an Apache-2.0 licensed base model.

Downloads last month: 10

Safetensors

Model size

4B params

Tensor type

F32

BF16