Python
DataNinja
Introduction DataNinja is a web application that supports to label data, design algorithms, train algorithms, and test algorithms. Related Patent CN111709522A: Deep learning target detection system based on server-embedded cooperation Code Repository The source code of the project is not publicly available. Screenshots
Compile Caffe without Root Privileges
In this tutorial, we are going to introduce how to install Caffe without root privileges. We assume that you have installed Anaconda and CUDA on your PC. Create Virtual Environment conda create -n caffe conda activate caffe Install Dependencies Since we decide not to use system dependencies, we need to install them in the Anaconda environment. To install, you can use the following commands: conda install boost=1.65.1 openblas mkl mkl-include gflags glog lmdb leveldb h5py hdf5 scikit-image conda install -c conda-forge ffmpeg opencv==3.4.3 Build Protocol Buffer (protobuf) Please DO NOT install it with Anaconda, because it causes the problem of the undefined reference in the linking process.
Similar Patient Finder
Introduction Similar Patient Finder is a user-friendly, web-based diagnosis tool that enables physicians to find similarities between patients at the molecular level. It provides diverse methods to process data and a clear visualization of patients’ correlation, allowing physicians to easily identify a patient’s disease subtype, stage of disease, morbid state, and provide appropriate therapy advice or guidance on prescriptions. Code Repository GitLab Screenshots
Making Requests Non-blocking in Tornado
Tornado is one of the most popular web frameworks for Python, which is based on a single thread IO loop (aka event loop). You can handle high concurrency with optimal performance. However, Tornado is single-threaded (in its common usage, although it supports multiple threads in advanced configurations), therefore any “blocking” task will block the whole server. This means that a blocking task will not allow the framework to pick the next task waiting to be processed.
Medical Image Tagger
Introduction Medical Image Tagger is a collaborative project between my team and Harvard Medical University. It was designed to improve the efficiency of medical students by 20% by using medical domain knowledge and a web service to tag medical images with relevant medical tags and image types. I was responsible for designing and developing the backend logic and database schema. The initial version of the project was built using Java, and after collecting nearly 1 million datasets from the web service, an offline function was added. The second version of the project was built using Python and Django.