Instructions to use poolside/Laguna-XS.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use poolside/Laguna-XS.2 with Transformers:

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

pipe = pipeline("text-generation", model="poolside/Laguna-XS.2", trust_remote_code=True)
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("poolside/Laguna-XS.2", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("poolside/Laguna-XS.2", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
AMD Developer Cloud
Local Apps Settings

vLLM

How to use poolside/Laguna-XS.2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "poolside/Laguna-XS.2"
# 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-XS.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/poolside/Laguna-XS.2

SGLang

How to use poolside/Laguna-XS.2 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-XS.2" \
    --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-XS.2",
		"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-XS.2" \
        --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-XS.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use poolside/Laguna-XS.2 with Docker Model Runner:
```
docker model run hf.co/poolside/Laguna-XS.2
```

Laguna-XS.2

Commit History

Aug 1

00:56
Align chat template `enable_thinking` default with generation_config.json and the model card⁠

1MrazorT1

Verified signature
3db49e1

Jul 14

08:34
fix(rope): YaRN attention_factor is (0.1*ln(factor)+1)*attn_factor, not the bare multiplier

joerowell

Verified signature
69e3f40

Jul 1

08:57
Use chat_template.jinja as the single source: drop the {% include %} chat_template field from tokenizer_config.json

joerowell

Verified signature
3ff0d45

Jun 18

18:34
small self-contained fixes

joerowell

Verified signature
f5f4885
09:46
Document SGLang support (sgl-project/sglang#24204)

joerowell

Verified signature
0bab366

Jun 3

12:33
Update README.md

varunrandery

Verified signature
0f4935b

Jun 1

18:30
Update README.md (#9)⁠

Tzachi1980 +1: baranowskiadam

53cd246

May 28

12:24
Update .eval_results/terminal-bench-2.0.yaml

joerowell

Verified signature
1162d99
12:24
Update .eval_results/swe-bench_verified.yaml

joerowell

Verified signature
c5fe6fd
12:23
Update .eval_results/swe-bench_pro.yaml

joerowell

Verified signature
3b9ed24

May 27

09:00
update sampling parameters to match evals

joerowell

07c6c8e

May 26

15:21
Update model card for 256K context length

varunrandery

Verified signature
a6da52e
14:59
increase context length to 256k

joerowell

339151d
14:57
Add tech report

joerowell

Verified signature
cfa46f7

May 20

20:12
Drop VLLM_USE_DEEP_GEMM=0 from vllm serve recipe (DeepGEMM is supported on Hopper and datacenter Blackwell)

joerowell

Verified signature
bb5158c

May 19

09:28
Update vLLM install instructions: 0.21.0+ supports Laguna XS, no nightly needed

joerowell

Verified signature
3c519e6

May 8

12:51
Add DFlash speculative decoding section

joerowell

e5916e1

May 6

18:58
Update README.md

joerowell

Verified signature
7ffaee5

May 3

12:56
Fix typo

Pengming

Verified signature
d38ff3c
12:54
Fix typo

Pengming

Verified signature
f07da4d

Apr 29

22:33
Enable thinking by default in recommended vllm serve command⁠

joerowell

Verified signature
be45e1e
22:27
Update vLLM serve command and link to vLLM recipes page⁠

joerowell

Verified signature
30ca10b
14:50
Drop install-from-source note for Transformers⁠

joerowell

Verified signature
de04ba5
14:46
Update vLLM install instructions to use nightly wheel⁠

joerowell

Verified signature
598f732

Apr 28

16:04
Add community evaluation results for SWE-BENCH_PRO, SWE-BENCH_VERIFIED, TERMINAL-BENCH-2.0 (#5)⁠

joerowell +1: nielsr

06367f7
14:18
Laguna XS.2 upload

joerowell

7b8e693

Commit History

Aug 1

Align chat template `enable_thinking` default with generation_config.json and the model card⁠

Jul 14

fix(rope): YaRN attention_factor is (0.1*ln(factor)+1)*attn_factor, not the bare multiplier

Jul 1

Use chat_template.jinja as the single source: drop the {% include %} chat_template field from tokenizer_config.json

Jun 18

small self-contained fixes

Document SGLang support (sgl-project/sglang#24204)

Jun 3

Update README.md

Jun 1

Update README.md (#9)⁠

May 28

Update .eval_results/terminal-bench-2.0.yaml

Update .eval_results/swe-bench_verified.yaml

Update .eval_results/swe-bench_pro.yaml

May 27

update sampling parameters to match evals

May 26

Update model card for 256K context length

increase context length to 256k

Add tech report

May 20

Drop VLLM_USE_DEEP_GEMM=0 from vllm serve recipe (DeepGEMM is supported on Hopper and datacenter Blackwell)

May 19

Update vLLM install instructions: 0.21.0+ supports Laguna XS, no nightly needed

May 8

Add DFlash speculative decoding section

May 6

Update README.md

May 3

Fix typo

Fix typo

Apr 29

Enable thinking by default in recommended vllm serve command⁠

Update vLLM serve command and link to vLLM recipes page⁠

Drop install-from-source note for Transformers⁠

Update vLLM install instructions to use nightly wheel⁠

Apr 28

Add community evaluation results for SWE-BENCH_PRO, SWE-BENCH_VERIFIED, TERMINAL-BENCH-2.0 (#5)⁠

Laguna XS.2 upload

fix(rope): YaRN attention_factor is (0.1ln(factor)+1)attn_factor, not the bare multiplier