Instructions to use thesumith/opensafety-1.5b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use thesumith/opensafety-1.5b with Transformers:

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

pipe = pipeline("text-generation", model="thesumith/opensafety-1.5b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("thesumith/opensafety-1.5b")
model = AutoModelForCausalLM.from_pretrained("thesumith/opensafety-1.5b", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use thesumith/opensafety-1.5b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "thesumith/opensafety-1.5b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "thesumith/opensafety-1.5b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/thesumith/opensafety-1.5b

SGLang

How to use thesumith/opensafety-1.5b 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 "thesumith/opensafety-1.5b" \
    --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": "thesumith/opensafety-1.5b",
		"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 "thesumith/opensafety-1.5b" \
        --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": "thesumith/opensafety-1.5b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use thesumith/opensafety-1.5b with Docker Model Runner:
```
docker model run hf.co/thesumith/opensafety-1.5b
```

🩺 OpenSafety-1.5B: Clinical Pharmacovigilance & MedDRA Coding Foundation Model

OpenSafety-1.5B is a specialized clinical language model trained for end-to-end Pharmacovigilance (PV). It extracts suspect drugs, identifies Adverse Drug Reactions (ADRs) from unstructured clinical text, maps symptoms to MedDRA v28.0 Preferred Terms (PT) and System Organ Classes (SOC), and performs regulatory Listedness (Expectedness) Assessment against Reference Safety Information (RSI).

This model powers the official opensafety Python package.

Model Details

Developed by: OpenSafety Team (thesumith)
Model Type: Causal Language Model (Fine-tuned Transformer)
Parameters: 1.5 Billion
Language: English
License: Apache 2.0
Supported Standards: MedDRA v28.0, ICH E2B(R3)
Python Library: opensafety on PyPI

Key Capabilities

Suspect Drug Identification: Identifies the primary suspect drug and dosage regimen from complex clinical narratives.
Adverse Event (AE) Extraction: Extracts raw patient verbatim descriptions.
MedDRA v28.0 Semantic Coding: Automatically maps patient language to official MedDRA Lowest Level Terms (LLT), Preferred Terms (PT), and Primary System Organ Classes (SOC).
Regulatory Listedness Assessment: Cross-references reported events against drug package inserts (FDA USPI, SmPC Section 4.8) to categorize events as LISTED or UNLISTED with regulatory audit justifications.

How to Use

Method 1: Using the Official `opensafety` Python Library (Recommended)

pip install opensafety

Downloads last month: 112

Safetensors

Model size

2B params

Tensor type

F32

F16

Model tree for thesumith/opensafety-1.5b