Instructions to use Parallel-Reasoning/llama-sosp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Parallel-Reasoning/llama-sosp with Transformers:

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

pipe = pipeline("text-generation", model="Parallel-Reasoning/llama-sosp")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Parallel-Reasoning/llama-sosp")
model = AutoModelForCausalLM.from_pretrained("Parallel-Reasoning/llama-sosp")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Parallel-Reasoning/llama-sosp with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Parallel-Reasoning/llama-sosp"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Parallel-Reasoning/llama-sosp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Parallel-Reasoning/llama-sosp

SGLang

How to use Parallel-Reasoning/llama-sosp 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 "Parallel-Reasoning/llama-sosp" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Parallel-Reasoning/llama-sosp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Parallel-Reasoning/llama-sosp" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Parallel-Reasoning/llama-sosp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Parallel-Reasoning/llama-sosp with Docker Model Runner:
```
docker model run hf.co/Parallel-Reasoning/llama-sosp
```

longlian

nielsr HF Staff commited on Apr 24, 2025

Commit

57f6521

verified ·

1 Parent(s): 5e058c3

Add model card (#1)

Browse files

- Add model card (5286a81e3e778dc7507164ad434c95377df51c50)

Co-authored-by: Niels Rogge <nielsr@users.noreply.huggingface.co>

Files changed (1) hide show

README.md +9 -0

README.md ADDED Viewed

	@@ -0,0 +1,9 @@

+---
+license: apache-2.0
+library_name: transformers
+pipeline_tag: text-generation
+---
+This repository contains the model described in the paper [Learning Adaptive Parallel Reasoning with Language Models](https://huggingface.co/papers/2504.15466).
+For more information, please refer to the github repository: https://github.com/Parallel-Reasoning/APR