metadata

tags:
  - merge
  - mergekit
  - lazymergekit
license: cc-by-nc-nd-4.0

JSL-MedMNX-7B

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "johnsnowlabs/JSL-MedMNX-7B"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

🏆 Evaluation

Tasks	Version	Filter	Metric	Value		Stderr
stem	N/A	none	acc_norm	0.5191	±	0.0068
		none	acc	0.5658	±	0.0058
- medmcqa	Yaml	none	acc	0.5135	±	0.0077
		none	acc_norm	0.5135	±	0.0077
- medqa_4options	Yaml	none	acc	0.5373	±	0.0140
		none	acc_norm	0.5373	±	0.0140
- anatomy (mmlu)	0	none	acc	0.6370	±	0.0415
- clinical_knowledge (mmlu)	0	none	acc	0.7245	±	0.0275
- college_biology (mmlu)	0	none	acc	0.7500	±	0.0362
- college_medicine (mmlu)	0	none	acc	0.6590	±	0.0361
- medical_genetics (mmlu)	0	none	acc	0.7200	±	0.0451
- professional_medicine (mmlu)	0	none	acc	0.7206	±	0.0273
- pubmedqa	1	none	acc	0.7720	±	0.0188

Groups	Version	Filter	n-shot	Metric	Value		Stderr
stem	N/A	none	0	acc_norm	0.5191	±	0.0068
		none	0	acc	0.5658	±	0.0058