RichardErkhov/johnsnowlabs_-_JSL-MedMNX-7B-SFT-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

JSL-MedMNX-7B-SFT - GGUF

• Model creator: https://huggingface.co/johnsnowlabs/
• Original model: https://huggingface.co/johnsnowlabs/JSL-MedMNX-7B-SFT/

Name	Quant method	Size
JSL-MedMNX-7B-SFT.Q2_K.gguf	Q2_K	2.53GB
JSL-MedMNX-7B-SFT.IQ3_XS.gguf	IQ3_XS	2.81GB
JSL-MedMNX-7B-SFT.IQ3_S.gguf	IQ3_S	2.96GB
JSL-MedMNX-7B-SFT.Q3_K_S.gguf	Q3_K_S	2.95GB
JSL-MedMNX-7B-SFT.IQ3_M.gguf	IQ3_M	3.06GB
JSL-MedMNX-7B-SFT.Q3_K.gguf	Q3_K	3.28GB
JSL-MedMNX-7B-SFT.Q3_K_M.gguf	Q3_K_M	3.28GB
JSL-MedMNX-7B-SFT.Q3_K_L.gguf	Q3_K_L	3.56GB
JSL-MedMNX-7B-SFT.IQ4_XS.gguf	IQ4_XS	3.67GB
JSL-MedMNX-7B-SFT.Q4_0.gguf	Q4_0	3.83GB
JSL-MedMNX-7B-SFT.IQ4_NL.gguf	IQ4_NL	3.87GB
JSL-MedMNX-7B-SFT.Q4_K_S.gguf	Q4_K_S	3.86GB
JSL-MedMNX-7B-SFT.Q4_K.gguf	Q4_K	4.07GB
JSL-MedMNX-7B-SFT.Q4_K_M.gguf	Q4_K_M	4.07GB
JSL-MedMNX-7B-SFT.Q4_1.gguf	Q4_1	4.24GB
JSL-MedMNX-7B-SFT.Q5_0.gguf	Q5_0	4.65GB
JSL-MedMNX-7B-SFT.Q5_K_S.gguf	Q5_K_S	4.65GB
JSL-MedMNX-7B-SFT.Q5_K.gguf	Q5_K	4.78GB
JSL-MedMNX-7B-SFT.Q5_K_M.gguf	Q5_K_M	4.78GB
JSL-MedMNX-7B-SFT.Q5_1.gguf	Q5_1	5.07GB
JSL-MedMNX-7B-SFT.Q6_K.gguf	Q6_K	5.53GB
JSL-MedMNX-7B-SFT.Q8_0.gguf	Q8_0	7.17GB

Original model description:

license: cc-by-nc-nd-4.0 language: - en library_name: transformers tags: - reward model - RLHF - medical

JSL-MedMNX-7B-SFT

JSL-MedMNX-7B-SFT is a 7 Billion parameter model developed by John Snow Labs.

This model is SFT-finetuned on alpaca format 11k medical dataset over the base model JSL-MedMNX-7B. Checkout the perofrmance on Open Medical LLM Leaderboard.

This model is available under a CC-BY-NC-ND license and must also conform to this Acceptable Use Policy. If you need to license this model for commercial use, please contact us at [email protected].

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "johnsnowlabs/JSL-MedMNX-7B-SFT"
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	n-shot	Metric	Value		Stderr
stem	N/A	none	0	acc_norm	0.5209	±	0.0068
		none	0	acc	0.5675	±	0.0058
- medmcqa	Yaml	none	0	acc	0.5152	±	0.0077
		none	0	acc_norm	0.5152	±	0.0077
- medqa_4options	Yaml	none	0	acc	0.5397	±	0.0140
		none	0	acc_norm	0.5397	±	0.0140
- anatomy (mmlu)	0	none	0	acc	0.6593	±	0.0409
- clinical_knowledge (mmlu)	0	none	0	acc	0.7245	±	0.0275
- college_biology (mmlu)	0	none	0	acc	0.7431	±	0.0365
- college_medicine (mmlu)	0	none	0	acc	0.6532	±	0.0363
- medical_genetics (mmlu)	0	none	0	acc	0.7300	±	0.0446
- professional_medicine (mmlu)	0	none	0	acc	0.7206	±	0.0273
- pubmedqa	1	none	0	acc	0.7720	±	0.0188

Groups	Version	Filter	n-shot	Metric	Value		Stderr
stem	N/A	none	0	acc_norm	0.5209	±	0.0068
		none	0	acc	0.5675	±	0.0058