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

Libraries

How to use poolside/Laguna-S-2.1 with Transformers:

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

pipe = pipeline("text-generation", model="poolside/Laguna-S-2.1", 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("poolside/Laguna-S-2.1", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("poolside/Laguna-S-2.1", 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 poolside/Laguna-S-2.1 with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/poolside/Laguna-S-2.1

SGLang

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

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

Is 1M context length available only on BF16 weights? Is FP8 constrained to 256k?

#25

by TempleOfApshai - opened about 20 hours ago

Discussion

TempleOfApshai

about 20 hours ago

Trying to load Laguna S 2.1 in FP8 weights defaults to 256k tokens context in vLLM. Forcing --max-model-len 1048576 throws an error. Using full BF16 safetensors immediately gives 1M context.

Is 1M only for the non-quantized versions?

TempleOfApshai

about 20 hours ago

Oh I think they're both 256k and the BF16 is (mis)configured to use RoPE:

(APIServer pid=672117) [transformers] Unrecognized keys in `rope_parameters` for 'rope_type'='default': {'full_attention', 'sliding_attention'}
(APIServer pid=672117) [transformers] Unrecognized keys in `rope_parameters` for 'rope_type'='default': {'full_attention', 'sliding_attention'}
(APIServer pid=672117) [transformers] The tokenizer you are loading from '/models/models--poolside--Laguna-S-2.1/snapshots/b0a9fd7c850eea9d68cf3ffc5bd5d3c5dced43c0' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.

Does this mean that Laguna S 2.1 isn't actually natively 1M and the claimed 1M context is really just 256k + RoPE?

DopplerFreq

about 19 hours ago

•

edited about 19 hours ago

The FP8 model card on HF describes how to enable the larger than 256k context on the quantized versions. “ The weights are native 1M checkpoints: training included a long-context extension stage up to 1,048,576 tokens, and quantization was calibrated at the 1M configuration. If you need more than 256K of context, restore the 1M configuration by editing config.json”. Different temperature and top_p parameters are also recommended. “ For the best balance of quality and reliability we recommend sampling with temperature 0.7 and top_p 0.95.” In general, I have seen more problems with open code when I followed those steps and tried to use 512k ctx. This was at BF16 KV too.

allenc87

about 4 hours ago

Are you using "VLLM_ALLOW_LONG_MAX_MODEL_LEN=1" in vLLM?

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