Instructions to use SupraLabs/Supra-Router-51M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SupraLabs/Supra-Router-51M with Transformers:

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

pipe = pipeline("text-generation", model="SupraLabs/Supra-Router-51M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SupraLabs/Supra-Router-51M")
model = AutoModelForCausalLM.from_pretrained("SupraLabs/Supra-Router-51M")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use SupraLabs/Supra-Router-51M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SupraLabs/Supra-Router-51M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SupraLabs/Supra-Router-51M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/SupraLabs/Supra-Router-51M

SGLang

How to use SupraLabs/Supra-Router-51M 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 "SupraLabs/Supra-Router-51M" \
    --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": "SupraLabs/Supra-Router-51M",
		"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 "SupraLabs/Supra-Router-51M" \
        --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": "SupraLabs/Supra-Router-51M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use SupraLabs/Supra-Router-51M with Docker Model Runner:
```
docker model run hf.co/SupraLabs/Supra-Router-51M
```

How did you determine the sequence length

by TimeLordRaps - opened 1 day ago

Discussion

TimeLordRaps

1 day ago

What other options of paths did you check, any with conditional branching?

QyrouNnet-AI

SupraLabs org about 19 hours ago

We did not do anything serious with this model.
@LH-Tech-AI made a simple model for our (and your apps) so that some could exist

This is mainly for SupraCode, but we are working on a better one 🤗😄

QyrouNnet-AI

SupraLabs org about 19 hours ago

PS: it can only accept one prompt, and it will generate the routing metadata!

QyrouNnet-AI changed discussion status to closed about 19 hours ago

roguexz

about 7 hours ago

•

edited about 7 hours ago

Do you have any plans on releasing a serious version of this model? Not sure why this user prompt failed.

summarize text
Supra-Router-51M is an ultra-lightweight, high-speed infrastructure traffic controller optimized for localized edge orchestration. 
With only 51.7 million parameters, this micro-LLM acts as a defensive gateway for multi-model ecosystems, accurately determining
when user requests can be processed locally by an Edge SLM or when they must be triaged to a cloud-hosted frontier intelligence layer.

The model was built by fine-tuning a pre-trained 51M base on the SupraLabs/Prompt-Routing-Dataset (992 rows). Rather than acting
as a naive binary classifier, the model uses Multi-Task Sequence Generation to map out the underlying properties of a prompt before
predicting the final routing token, anchoring its attention heads to robust language and structural logic features.

Domain: Content Creation | Complexity: 4 | Math: False | Code: True | Route: big model | Justification: Automated override: Task complexity is high (4) or involves technical logic (Math: False, Code: True), requiring frontier capabilities.

I mean, edge models should be able to handle that.

AxionLab-official

SupraLabs org about 7 hours ago

Do you have any plans on releasing a serious version of this model? Not sure why this user prompt failed.

wdym serious version lol

roguexz

about 4 hours ago

wdym serious version lol

@AxionLab-Official Well I was reading the above comment which made it look like this model is basically a toy and not ready for primetime.

We did not do anything serious with this model.
@LH-Tech-AI made a simple model for our (and your apps) so that some could exist

Also take a look at the analysis that I did. The classification seems a bit off.

TimeLordRaps

about 2 hours ago

I would say that this model is conceptually a great idea

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