Instructions to use AtomicChat/Laguna-S-2.1-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AtomicChat/Laguna-S-2.1-NVFP4 with Transformers:

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

pipe = pipeline("text-generation", model="AtomicChat/Laguna-S-2.1-NVFP4", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("AtomicChat/Laguna-S-2.1-NVFP4", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("AtomicChat/Laguna-S-2.1-NVFP4", trust_remote_code=True, 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use AtomicChat/Laguna-S-2.1-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AtomicChat/Laguna-S-2.1-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AtomicChat/Laguna-S-2.1-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/AtomicChat/Laguna-S-2.1-NVFP4

SGLang

How to use AtomicChat/Laguna-S-2.1-NVFP4 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 "AtomicChat/Laguna-S-2.1-NVFP4" \
    --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": "AtomicChat/Laguna-S-2.1-NVFP4",
		"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 "AtomicChat/Laguna-S-2.1-NVFP4" \
        --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": "AtomicChat/Laguna-S-2.1-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use AtomicChat/Laguna-S-2.1-NVFP4 with Docker Model Runner:
```
docker model run hf.co/AtomicChat/Laguna-S-2.1-NVFP4
```

Base model: poolside/Laguna-S-2.1

Laguna S 2.1, self-quantized to NVFP4 by Atomic Chat. Built straight from Poolside's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline.

Highlights

117.6B parameters: the weights this repo quantizes.
Context length: 1,048,576 tokens (1M), as published by Poolside.
48 layers: Mixture-of-Experts, hybrid sliding-window (512) and global attention.
Full imatrix ladder: every quant is calibrated with an importance matrix.
Mixed SWA and global attention layout: 48 layers in a 1:3 global-to-SWA ratio (12 global attention layers, 36 sliding-window layers, window 512), with softplus attention gating and per-layer-type rotary scales.
Native reasoning support: interleaved thinking between tool calls, with per-request control via enable_thinking.
Speculative decoding: a trained DFlash draft model is available for lower-latency serving.

These NVFP4s are self-quantized from the original weights, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.

Model Overview

Property	Value
Base model	`poolside/Laguna-S-2.1`
Parameters	117.6B
Layers	48
Experts	256 routed (top-10)
Sliding window	512 tokens
Context length	1,048,576 tokens (1M)
Vocabulary	100,352
Modalities	Text
Architecture	Mixture-of-Experts, 256 experts (top-10), hybrid sliding-window (512) and global attention, 48 attention heads over 8 KV heads, `LagunaForCausalLM`
This repo	NVFP4 weights

Benchmarks

Benchmark	Score
Laguna S 2.1	70.2%
Tencent Hy3	71.7%
Inkling	63.8%
Nemotron 3 Ultra	56.4%
DeepSeek-V4-Pro Max	64.0%
Kimi K3	88.3%
Qwen 3.7 Max	74.5%
Muse Spark 1.1	80%
Claude Fable 5	88%

Scores are Poolside's published results for the base poolside/Laguna-S-2.1, not our own measurements. Quantization preserves the large majority of this; Q4_K_M and up stay close to full precision.

Get started

Atomic Chat: search AtomicChat/Laguna-S-2.1-NVFP4 and hit Use this model.
vLLM: vllm serve AtomicChat/Laguna-S-2.1-NVFP4 --max-model-len 8192

Best practices

Parameter	Value
temperature	1.0
top_p	1.0
top_k	20
min_p	0.0

Poolside's recommended sampling configuration for poolside/Laguna-S-2.1.

How these were made

Download poolside/Laguna-S-2.1 (original weights).
Quantize to NVFP4 with llm-compressor over a calibration corpus.

License

Original model by Poolside, released under the OpenMDW-1.1 license. Full terms: OpenMDW-1.1. Quantized by Atomic Chat.

Downloads last month: 196

Safetensors

Model size

66B params

Tensor type

F32

BF16

F8_E4M3

Model tree for AtomicChat/Laguna-S-2.1-NVFP4

Base model

poolside/Laguna-S-2.1

Quantized

(70)

this model

Collection including AtomicChat/Laguna-S-2.1-NVFP4

Laguna S 2.1

Collection

4 items • Updated 4 days ago