Llama3-Preferred-MedSwallow-70B Open-source Medical Fine-tuned Model - Empowering Applications in the Field of Medical Research

Llama3 Preferred MedSwallow 70B

Developed by pfnet

Llama3-Preferred-MedSwallow-70B is a fine - tuned model that undergoes continuous pre - training based on the original corpus of medical - related texts and is specifically designed for the medical research field.

Large Language Model

Transformers

Supports Multiple Languages#Medical license exam #Japanese medical Q&A #70B large model

Downloads 295

Release Time : 7/17/2024

Model Overview

This model is a fine - tuned version based on Llama - 3 - Swallow - 70B - v0.1, focusing on text processing in the medical field and aiming to provide support for medical research.

Model Features

Optimized for the medical field

Underwent continuous pre - training on medical - related texts, improving performance in the medical field.

High performance

Performed excellently in the Japanese national medical license exam, outperforming mainstream models such as GPT - 4.

Large parameter scale

With 70B parameters, it has powerful text processing capabilities.

Model Capabilities

Medical text understanding

Medical Q&A

Medical knowledge reasoning

Use Cases

Medical education

Medical license exam assistance

Used to assist medical students in preparing for the national medical license exam.

Achieved an average score of 395.2 in the Japanese national medical license exam from 2018 - 2022, outperforming GPT - 4.

Medical research

Medical literature analysis

Helps researchers quickly understand and analyze medical literature.

🚀 Llama3-Preferred-MedSwallow-70B

Llama3-Preferred-MedSwallow-70B is a fine - tuned medical - related language model, offering high - quality text generation capabilities.

🚀 Quick Start

If you want to know more about the Llama3 - Preferred - MedSwallow - 70B model, you can refer to the official blog post at https://tech.preferred.jp/ja/blog/llama3-preferred-medswallow-70b/.

✨ Features

Fine - tuned on medical text: Llama3 - Preferred - MedSwallow - 70B is based on tokyotech - llm/Llama - 3 - Swallow - 70B - v0.1 and has undergone continued pre - training on a medical - related text corpus.
High performance: As shown in the performance table, it outperforms many other models in the Japanese national medical licensing examinations from 2018 to 2022.

📚 Documentation

Model Description

Llama3 - Preferred - MedSwallow - 70B is a finetuned model based on tokyotech - llm/Llama - 3 - Swallow - 70B - v0.1, which has undergone continued pretraining on an original corpus of medical - related text. For more details, please refer to our blog post at https://tech.preferred.jp/ja/blog/llama3-preferred-medswallow-70b/. The model is released under the META LLAMA 3 COMMUNITY LICENSE.

Model Performance

The table below shows the performance on the Japanese national medical licensing examinations from 2018 to 2022 (IgakuQA).

Model ID	Average	2018	2019	2020	2021	2022
Llama3 - Preferred - MedSwallow - 70B	395.2	407	390	391	393	395
GPT - 4	388.8	382	385	387	398	392
Llama - 3 - Swallow - 70B - v0.1	348.6	353	347	353	345	345
Meta - Llama - 3 - 70B	334.6	353	340	348	314	318
Qwen2 - 72B	331.2	320	325	325	326	360
gemma - 2 - 27b	316	337	298	327	296	322
Swallow - 70b - NVE - hf	291.6	283	280	300	295	300
Swallow - MX - 8x7b - NVE - v0.1	280.8	262	273	291	284	294
ChatGPT	273.2	266	250	266	297	287

Limitations

The model was developed for research purposes and is not intended for clinical diagnosis. It is the users' responsibility to ensure compliance with applicable rules and regulations.

Contributors

Preferred Networks, Inc.

Junichiro Iwasawa
Keita Suzuki
Wataru Kawakami

📄 License

META LLAMA 3 COMMUNITY LICENSE

Property	Details
Model Type	Llama
License	META LLAMA 3 COMMUNITY LICENSE
Pipeline Tag	text - generation
Library Name	transformers

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご