Instructions to use BrainHealthAI/BrainMed-8B-SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use BrainHealthAI/BrainMed-8B-SFT with Transformers:

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

pipe = pipeline("text-generation", model="BrainHealthAI/BrainMed-8B-SFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("BrainHealthAI/BrainMed-8B-SFT")
model = AutoModelForCausalLM.from_pretrained("BrainHealthAI/BrainMed-8B-SFT", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use BrainHealthAI/BrainMed-8B-SFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "BrainHealthAI/BrainMed-8B-SFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BrainHealthAI/BrainMed-8B-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/BrainHealthAI/BrainMed-8B-SFT

SGLang

How to use BrainHealthAI/BrainMed-8B-SFT 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 "BrainHealthAI/BrainMed-8B-SFT" \
    --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": "BrainHealthAI/BrainMed-8B-SFT",
		"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 "BrainHealthAI/BrainMed-8B-SFT" \
        --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": "BrainHealthAI/BrainMed-8B-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use BrainHealthAI/BrainMed-8B-SFT with Docker Model Runner:
```
docker model run hf.co/BrainHealthAI/BrainMed-8B-SFT
```

BrainMed-8B-SFT / evaluation /SIGNIFICANCE.md

Williamsanderson

evaluation results and figures

8a33135 verified 16 days ago

preview code

Raw

History Blame Contribute Delete

1.48 kB

Paired significance: `brainmed-8b-v1__soup-last-a0.3` vs `base-HuatuoGPT-o1-8B`

McNemar's exact test on the items both runs answered. only B counts items the candidate gets right and the baseline gets wrong; only A the reverse. Items both get right or both get wrong are uninformative about which model is better.

Benchmark	n	base-HuatuoGPT-o1-8B	brainmed-8b-v1__soup-last-a0.3	Delta	only A	only B	p	verdict (a=0.05)
MB-op4	298	59.73	62.42	+2.68	30	38	0.396	no difference
MB-op5	298	55.37	53.36	-2.01	37	31	0.545	no difference
MedMCQA	4183	63.28	64.04	+0.77	358	390	0.257	no difference
MedQA	1273	77.93	76.59	-1.34	103	86	0.244	no difference
MedXpert	1449	16.98	18.63	+1.66	93	117	0.112	no difference
PubMedQA	1000	80.2	79.1	-1.10	30	19	0.152	no difference

Pooled across benchmarks

items only base-HuatuoGPT-o1-8B gets right: 651
items only brainmed-8b-v1__soup-last-a0.3 gets right: 681
McNemar exact p = 0.4269

Pooled, the two runs are statistically indistinguishable (p = 0.427). Any average difference between them is within sampling noise.

Per benchmark: 0 significantly better, 0 significantly worse, 6 indistinguishable.

Pooling treats benchmarks as one sample and so weights them by size; read it alongside the per-benchmark rows, not instead of them.

Paired significance: brainmed-8b-v1__soup-last-a0.3 vs base-HuatuoGPT-o1-8B

Pooled across benchmarks

Paired significance: `brainmed-8b-v1__soup-last-a0.3` vs `base-HuatuoGPT-o1-8B`