Inspiration

It is time consuming to take down notes and write a to-do list after every meeting. It is also a bit distracting to write while listening. What if this task can be automated?

What it does

This tool summarizes meeting transcript and identify action items to automatically create a to-do list.

How we built it

  1. Train summarizer and action item classifier.
  2. Launch AWS EC2 DL1 instance.
  3. Connect to AWS EC2 DL1 instance.

    ssh -i /path/my-key-pair.pem my-instance-user-name@my-instance-public-dns-name
    
  4. Migrate models and scripts to EC2 via Filezilla.

  5. Pull Docker Image - Habana Vault

    docker pull vault.habana.ai/gaudi-docker/1.3.0/{$OS}/habanalabs/tensorflow-installer-tf-cpu-${TF_VERSION}:1.3.0-499
    
  6. Run Docker to mount repo

    docker run -v ${repo_path_in_EC2}:/repo -it --runtime=habana -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --net=host vault.habana.ai/gaudi-docker/1.3.0/{$OS}/habanalabs/tensorflow-installer-tf-cpu-${TF_VERSION}:1.3.0-499
    
  7. Install dependencies

    pip install -r requirements.txt --quiet
    
  8. Run

    cd repo
    python3 src/main.py
    

Challenges we ran into

  • Unfamiliarity with AWS DL1.
  • Difficulty finding relevant open source datasets

Accomplishments that we're proud of

  • We were able to use AWS EC2 DL1 for the first time!
  • We were able to build our first summarizer model!
  • We were able to build our first action item classifier!

What we learned

  • We learned how to use AWS EC2 DL1!
  • We learned more about NLP!

What's next for ToDo List Automation

  • Integration with teleconference apps (Zoom, MS teams, etc…)
  • Improve performance of summarizer and action item classifier
  • Optimize pipeline using other AWS services
Share this project:

Updates