What it does

This repo collects links, talks, and practical advice for surviving a computer vision PhD. Topics span research conduct, paper reading, writing for CVPR-tier venues, reviewing, code release, neural network tricks, and coding tips. Each section lives in its own markdown file with short summaries and quoted highlights tucked into dropdowns.

The interesting bit

The author started compiling this in 2017 after repeatedly discovering resources they wished they’d found earlier. The explicit goal is scannability: overviews in under a minute, original sources always linked. It’s a personal notebook that happens to be public.

Key highlights

• Research advice, paper reading strategies, and writing guides each get dedicated curated lists
• Code release guide covers making work “easy to understand and to reuse”
• Neural network tricks and coding tips sections included for practical implementation hurdles
• Summaries often quote original sources directly, with extensive use of <details> dropdowns for brevity
• Open to contributions via pull requests and issues

Caveats

• Three planned sections remain unchecked since 2017: fast/reproducible experimentation, supervisor meetings, and beautiful visualizations
• Scope is explicitly personal and computer-vision-centric; other fields may find limited relevance

Verdict

Worth a bookmark if you’re early in a vision PhD or about to start one. Established researchers probably already have their own versions of these lists.