Instructions to use SNOWTEAM/MedicoLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SNOWTEAM/MedicoLLM with Transformers:

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

pipe = pipeline("text-generation", model="SNOWTEAM/MedicoLLM")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SNOWTEAM/MedicoLLM")
model = AutoModelForCausalLM.from_pretrained("SNOWTEAM/MedicoLLM")
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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use SNOWTEAM/MedicoLLM with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SNOWTEAM/MedicoLLM"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SNOWTEAM/MedicoLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SNOWTEAM/MedicoLLM

SGLang

How to use SNOWTEAM/MedicoLLM 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 "SNOWTEAM/MedicoLLM" \
    --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": "SNOWTEAM/MedicoLLM",
		"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 "SNOWTEAM/MedicoLLM" \
        --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": "SNOWTEAM/MedicoLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use SNOWTEAM/MedicoLLM with Docker Model Runner:
```
docker model run hf.co/SNOWTEAM/MedicoLLM
```

Overview

SNOWTEAM/medico-mistral is a specialized language model designed for medical applications. This transformer-based decoder-only language model is based on the Mistral 8x7B model and has been fine-tuned through global parameter adjustments, leveraging a comprehensive dataset that includes 4.8 million research papers and 10,000 medical books.

Model Description

Base Model: Mistral 8x7B model- Instruct
Model type: Transformer-based decoder-only language model
Language(s) (NLP): English

Training Dataset

Dataset Size: 4.8 million research papers and 10,000 medical books.
Data Diversity: Includes a wide range of medical fields, ensuring comprehensive coverage of medical knowledge.
Preprocessing:
Books: We collected 10,000 textbooks from various sources such as the open-library, university libraries, and reputable publishers, covering a wide range of medical specialties. For preprocessing, we extracted text content from PDF files, then performed data cleaning through de-duplication and content filtering. This involved removing extraneous elements such as URLs, author lists, superfluous information, document contents, references, and citations.
Papers: Academic papers are a valuable knowledge resource due to their high-quality, cutting-edge medical information. We started with the S2ORC (Lo et al. 2020) dataset, which contains 81.1 million English-language academic papers. From this, we selected biomedical-related papers based on the presence of corresponding PubMed Central (PMC) IDs. This resulted in approximately 4.8 million biomedical papers, totaling over 75 billion tokens.

Model Sources [optional]

Repository: https://huggingface.co/SNOWTEAM/medico-mistral
Paper [optional]:
Demo [optional]:

How to Get Started with the Model

import transformers
import torch

model_path = "SNOWTEAM/medico-mistral"
model = AutoModelForCausalLM.from_pretrained(
    model_path,device_map="auto", 
    max_memory=max_memory_mapping,
    torch_dtype=torch.float16,
)
tokenizer = AutoTokenizer.from_pretrained("SNOWTEAM/medico-mistral")
input_text = ""
input_ids = tokenizer(input_text, return_tensors="pt").input_ids
output_ids = model.generate(input_ids=input_ids.cuda(),
                            max_new_tokens=300,
                            pad_token_id=tokenizer.eos_token_id,)
output_text = tokenizer.batch_decode(output_ids[:, input_ids.shape[1]:],skip_special_tokens=True)[0]
print(output_text)

Training Details

Training Hyperparameters

Training regime: [More Information Needed]

Speeds, Sizes, Times [optional]

Evaluation

Testing Data, Factors & Metrics

Testing Data

[More Information Needed]

Factors

[More Information Needed]

Metrics

[More Information Needed]

Results

[More Information Needed]

Summary

Citation [optional]

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]

Downloads last month: 11