Instructions to use EpistemeAI/Reasoning-Medical-27B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use EpistemeAI/Reasoning-Medical-27B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="EpistemeAI/Reasoning-Medical-27B") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("EpistemeAI/Reasoning-Medical-27B") model = AutoModelForMultimodalLM.from_pretrained("EpistemeAI/Reasoning-Medical-27B", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use EpistemeAI/Reasoning-Medical-27B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "EpistemeAI/Reasoning-Medical-27B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "EpistemeAI/Reasoning-Medical-27B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/EpistemeAI/Reasoning-Medical-27B
- SGLang
How to use EpistemeAI/Reasoning-Medical-27B 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 "EpistemeAI/Reasoning-Medical-27B" \ --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": "EpistemeAI/Reasoning-Medical-27B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'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 "EpistemeAI/Reasoning-Medical-27B" \ --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": "EpistemeAI/Reasoning-Medical-27B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Unsloth Studio
How to use EpistemeAI/Reasoning-Medical-27B with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for EpistemeAI/Reasoning-Medical-27B to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for EpistemeAI/Reasoning-Medical-27B to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for EpistemeAI/Reasoning-Medical-27B to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="EpistemeAI/Reasoning-Medical-27B", max_seq_length=2048, ) - Docker Model Runner
How to use EpistemeAI/Reasoning-Medical-27B with Docker Model Runner:
docker model run hf.co/EpistemeAI/Reasoning-Medical-27B
Might be even better with a properly abliterated model
This is just a suggestion. Using a properly abliterated model might be better for this purpose than vanilla corporate gimp that you used.
Something like: https://huggingface.co/llmfan46/Qwen3.6-27B-uncensored-heretic-v2
This one is really good, it maintained its intelligence and is WAY less irritating with it spineless disclaimers and fake concerns. (ignore the stupid gif on its page, the model is amazing.)
Hi, Sorry, we can't do that. It is uncensored and it will cause misalignment and will be less safe. I want this Medical Reasoning to be aligned and safe to use.
Thomas
As you like, but you should actually test it before writing it off at this early stage. Don't fine-tune yet, just download it and try it.
Most people don't actually understand what an uncensored model is. There is a damn good reason the Swiss Federal Supreme Court turned to Heretic abliteration instead of the neutered, condescending corporate drone vanillas. Medicine, just like the legal system, deals with raw reality. It is way too sensitive to function properly in all its nuances when bound by a corporate HR cancer mindset.
Don't worry, an abliterated model isn't wild or feral. It doesn't start spewing garbage unless a user explicitly forces it to. It's simply freed of artificial rigidity and fakery (to a limited extent). Underneath it all, the fundamental drive to be a helpful, precise assistant is still entirely intact, it just no longer panics when exposed to sensitive, real-world facts.
It might sound shocking but its literally more wholesome compared to default HR version, in every way.
Medicine, just like the legal system, deals with raw reality.
So does a murderer.
In many areas of the world, the idea has taken hold that the manufacturer of something is legally liable for how the thing is used. Many lawsuits have been filed and awarded based on this problematic view of criminal liability.
If there is a nonzero risk for the authors of a LLM to get dragged in front of a magistrate just because someone prompted their model for how to commit a poisoning, those authors may conclude this risk outweighs the $0.00 profit to be made from publishing an abliterated model.
The only solution to this is to do battle in the spiritual realm : reject and promote rejection of the idea that the creator of the product assumes the responsibility naturally accruing to the user of the product.
Thanks for the model EpisteLabs. If you get around to some description or documentation of your tuning and scaffolding i'm sure some people would be interested.