Awesome

MedJamba

Multilingual Medical Model Based On Jamba

📃 <a href="https://arxiv.org/abs/2403.03640" target="_blank">Paper</a> • 🌐 <a href="https://apollo.llmzoo.com/" target="_blank">Demo</a> • 🤗 <a href="https://huggingface.co/datasets/FreedomIntelligence/ApolloCorpus" target="_blank">ApolloCorpus</a> • 🤗 <a href="https://huggingface.co/datasets/FreedomIntelligence/XMedbench" target="_blank">XMedBench</a>

🌈 Update

• [2024.04.25] MedJamba Model is published！🎉

Results

🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-0.5B" target="_blank">Apollo-0.5B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-1.8B" target="_blank">Apollo-1.8B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-2B" target="_blank">Apollo-2B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-6B" target="_blank">Apollo-6B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-7B" target="_blank">Apollo-7B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-34B" target="_blank">Apollo-34B</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-72B" target="_blank">Apollo-72B</a>

🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-MedJamba" target="_blank">MedJamba</a>

🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-0.5B-GGUF" target="_blank">Apollo-0.5B-GGUF</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-2B-GGUF" target="_blank">Apollo-2B-GGUF</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-6B-GGUF" target="_blank">Apollo-6B-GGUF</a> • 🤗 <a href="https://huggingface.co/FreedomIntelligence/Apollo-7B-GGUF" target="_blank">Apollo-7B-GGUF</a>

Dataset & Evaluation

• Dataset 🤗 <a href="https://huggingface.co/datasets/FreedomIntelligence/ApolloCorpus" target="_blank">ApolloCorpus

  <details><summary>Click to expand</summary>

  

  • Zip File
  • Data category
    • Pretrain:
      • data item:
        • json_name: {data_source}{language}{data_type}.json
        • data_type: medicalBook, medicalGuideline, medicalPaper, medicalWeb(from online forum), medicalWiki
        • language: en(English), zh(chinese), es(spanish), fr(french), hi(Hindi)
        • data_type: qa(generated qa from text)
        • data_type==text: list of string

          [
            "string1",
            "string2",
            ...
          ]
          

        • data_type==qa: list of qa pairs(list of string)

          [
            [
              "q1",
              "a1",
              "q2",
              "a2",
              ...
            ],
            ...
          ]
          

    • SFT:
      • json_name: {data_source}_{language}.json
      • data_type: code, general, math, medicalExam, medicalPatient
      • data item: list of qa pairs(list of string)

          [
            [
              "q1",
              "a1",
              "q2",
              "a2",
              ...
            ],
            ...
          ]
        

  </details>

• Evaluation 🤗 <a href="https://huggingface.co/datasets/FreedomIntelligence/XMedbench" target="_blank">XMedBench</a>

  <details><summary>Click to expand</summary>

  • EN:

    • MedQA-USMLE
    • MedMCQA
    • PubMedQA: Because the results fluctuated too much, they were not used in the paper.
    • MMLU-Medical
      • Clinical knowledge, Medical genetics, Anatomy, Professional medicine, College biology, College medicine

  • ZH:

    • MedQA-MCMLE
    • CMB-single: Not used in the paper
      • Randomly sample 2,000 multiple-choice questions with single answer.
    • CMMLU-Medical
      • Anatomy, Clinical_knowledge, College_medicine, Genetics, Nutrition, Traditional_chinese_medicine, Virology
    • CExam: Not used in the paper
      • Randomly sample 2,000 multiple-choice questions

  • ES: Head_qa

  • FR: Frenchmedmcqa

  • HI: MMLU_HI

    • Clinical knowledge, Medical genetics, Anatomy, Professional medicine, College biology, College medicine

  • AR: MMLU_Ara

    • Clinical knowledge, Medical genetics, Anatomy, Professional medicine, College biology, College medicine
  </details>

Results reproduction

1. Download Dataset for project:

   bash 0.download_data.sh
   

2. Prepare test and dev for specific model:

   • Create test data for with special token, you can use ./util/check.ipynb to check models' special tokens

   bash 1.data_process_test&dev.sh
   

3. Prepare train data for specific model (Create tokenized data in advance):

   • You can adjust data Training order and Training Epoch in this step

   bash 2.data_process_train.sh
   

4. Train the model

   • Multi Nodes refer to ./scripts/multi_node_train_*.sh

   pip install causal-conv1d>=1.2.0
   pip install mamba-ssm
   

   Node 0:

   bash ./scripts/3.multinode_train_jamba_rank0.sh
   

   ... Node 4:

   bash ./scripts/3.multinode_train_jamba_rank4.sh
   

5. Evaluate your model: Generate score for benchmark

   bash 4.eval.sh
   

6. Evaluate your model: Play with your ckpts in bash

   python ./src/evaluate/cli_demo.py --model_name='./ckpts/your/path/tfmr'
   

To do

• Long Context Capability Evaluation and new Long-Med Benchmark

Acknowledgment

• HuatuoGPT-II
• proxy-tuning
• Apollo

Citation

Please use the following citation if you intend to use our dataset for training or evaluation:

@misc{wang2024apollo,
   title={Apollo: Lightweight Multilingual Medical LLMs towards Democratizing Medical AI to 6B People},
   author={Xidong Wang and Nuo Chen and Junyin Chen and Yan Hu and Yidong Wang and Xiangbo Wu and Anningzhe Gao and Xiang Wan and Haizhou Li and Benyou Wang},
   year={2024},
   eprint={2403.03640},
   archivePrefix={arXiv},
   primaryClass={cs.CL}
}