Instructions to use Soofi-Project/Soofi-S-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Soofi-Project/Soofi-S-Base with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Soofi-Project/Soofi-S-Base") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Soofi-Project/Soofi-S-Base", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Soofi-Project/Soofi-S-Base with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Soofi-Project/Soofi-S-Base" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Soofi-Project/Soofi-S-Base", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Soofi-Project/Soofi-S-Base
- SGLang
How to use Soofi-Project/Soofi-S-Base 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 "Soofi-Project/Soofi-S-Base" \ --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": "Soofi-Project/Soofi-S-Base", "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 "Soofi-Project/Soofi-S-Base" \ --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": "Soofi-Project/Soofi-S-Base", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Soofi-Project/Soofi-S-Base with Docker Model Runner:
docker model run hf.co/Soofi-Project/Soofi-S-Base
Typisch deutsch, alles gaaaaaaanz langsam
Da liest man, freut sich und da steht der Türsteher und sagt du kommst hier nicht rein. Was sollte man auch erwarten wenn irgendwo deutsch und Telekom in einem Artikel gleichzeitig stehen. Das man sich einfach so was runterladen kann.
Ach nein, wie auch, das wäre nicht deutsch.
Deutsch ist das man erstmal beantragen muss, dann auf eine ellenlange warteliste kommt und bis man das Modell runterladen kann hat es der Chinese kopiert, verbessert, verteilt, im einsatz und Soofi verschwindet wie jede deutsche innovation in der belanglosigkeit....
Da hab ich schon gar kein bock mehr mich mit abzuggeben.
traurig aber wahr :-(
I'm really confused with the licensing model. In the arxiv paper, https://arxiv.org/abs/2607.09424 it's titled "A Sovereign, Open-Source Foundation Model for German and English" but the license in Huggingface is clearly non open-source. Can you clarify this? Thanks a lot.
Also whats happening with my Data, DSGVO?
Why a german ministry which is payed by my taxes needs my personal information to get access for a public model?
Why a model created from absolute scratch is a NVIDIA Nemotron-3 Finetune.
Where is everyone? Fleeing from the justice cause of scam? Everyone drunk of doing Party?
I dunnow whats happening but in the meanwhile the community is watching carefully and they have many questions.
I can't even post the responses i got posting this repository.
If you want feedback and don't fear it make it public and not a selection of company's without real knowledge.
Thats always the german problem. Bureaucracy. Let it go, let it flow, make it good make use of the community not war against.
"Why a model created from absolute scratch is a NVIDIA Nemotron-3 Finetune."
It is no finetune. It trained from scratch with 25T tokens. They just reused the architecture.
I think you are a bit exaggerating here.
Its not unlikely that this model will fade away quickly, because its gated (unfortuantely), but your statemetns are polemical.
Nemotron-3-a3b-30b is dumb—not because Nvidia is too incompetent to train it, but because the base model itself is dumb. It exists, but it has no real relevance. That isn't polemics; that’s reality.
Sagt mal, wie lange dauert denn das, bis man die Freigabe bekommt? Die AI Welt lebt von Geschwindigkeit, ist das bei der DT schon angekommen? Wenn das Modell schon wieder ein alter Hut ist, brauche ich auch keine Freigabe mehr.
Sagt mal, wie lange dauert denn das, bis man die Freigabe bekommt? Die AI Welt lebt von Geschwindigkeit, ist das bei der DT schon angekommen? Wenn das Modell schon wieder ein alter Hut ist, brauche ich auch keine Freigabe mehr.
Chill mal, die haben das hochgeladen, ihre Timeline erreicht. Jetzt sind die erstmal mit Party, Palmen, Weibern und nem Bier beschäftigt oder mit selbstbeweihräucherung oder so, keine Zeit für die Community, die sind alle viel zu busy um das hier zu lesen. 🤣
Könnte natürlich auch sein das die ne Soofi AI am laufen haben, die auf den freischalten Button drücken soll. 🤔
Also waiting for access for days now. Even with Meta's gated stuff access was granted immediately?
What checks do they make on my Email-address that takes days?
Even more annoying as the model seems to be available via LMStudio, so why not here in HF?
Edit model seem to have been withdrawn in LM_Studio, but still listed with 57 downloads?
What kind of souvereignty shall be demonstrated by such an intransparenter and murky publishing process?
They changed their mind... CLOSED BETA NOW. Seems that they have...
We hear you, and we truly appreciate the incredible interest in Soofi S.
The number of requests for early access, testing, and technical details has exceeded our expectations. We know many of you are eager to get your hands on the model.
The models on HuggingFace are intermediate results and in a closed beta with already selected partners to work on an improved version. Soofi S is currently in the final post-training stage. We are working hard to release a model that is not only open, but also robust, reliable, and genuinely useful for European businesses. Our final checkpoints will be freely available (with a permissive license), without any gated access.
Our target is an open release within the next weeks and will share access details as soon as the model is ready.
Thank you so much for your enthusiasm, your patience, and your support.