Skip to content

Latest commit

 

History

History
24 lines (24 loc) · 983 Bytes

README.md

File metadata and controls

24 lines (24 loc) · 983 Bytes

serve-healthcare

Create Conda Enviornment

  1. conda create -n serve-health python=3.6
  2. conda activate serve-health
  3. conda install -c pytorch pytorch
  4. conda install -c pytorch torchvision

Install Golang

Installation details can be found here.

Cloning the repository

git clone --recurse-submodules https://github.com/alindkhare/serve-healthcare.git

Installing Ray

  1. cd serve-healthcare
  2. bash ./ray/ci/travis/install-bazel.sh
  3. cd ./ray/python && pip install -e .[serve] --verbose
  4. cd ../../

Installing Python Packages

  1. pip install -r requirement.txt

Installing Profiling library library

  1. pip install -e . --verbose

Running the profiling code

Running python profile.py populates .jsonl file for time taken by each query. Using .jsonl file following can be plotted - Latency CDF Query Timeline