CS 425 Distributed Machine Learning Cluster(Best Rust MP * 4)

Getting started

We are (proudly) using Rust! You can use rustup to setup your Rust environment easily.

Usage

$ cargo run

Available CLI commands:

• list_mem, lm: list current membership list (only active members are shown)
• list_self: print current node's ID
• join, j: join the group
• leave, l: leave the group
• p[ut] <local_file_path> <remote_filename>: store a file in the file system
• g[et] <remote_filename> <local_file_path>: retrieve a file from the file system
• d[elete] <remote_filename>: delete a file from the file system
• ls <remote_filename>: list where a file is stored in the file system
• s[tore]: list all files stored in the current node
• get-versions/gv <remote_filename> <num_versions> <local_file_path>: get the last num_versions versions of a file
• t[rain]: train the machine learning models, in this case is loading the pretrained models
• predict: perform distributed inference on imagenet_1k
• jobs: See the current status of the prediction jobs, including percentiles

The program logs to HOSTNAME.log.

Report

Report is located here.

The data screenshots are located here.