metrics.help
Machine learning training metrics explained
3 followers
Machine learning training metrics explained
3 followers
Machine learning metrics and algorithms, explained. A simple math-free resource to learn about the different ML training metrics, including: - Simple definitions and explanations - Visualizations for healthy and unhealthy progressions - Deep dive into reinforcement learning algorithms
KL Divergance? Gradient Norm? Clip Ratio? I keep forgetting what these ML metrics mean and what to watch for in my LLM training runs.
Since I couldn’t find a single resource that explained them without heavy math, I created one:
👋 say hello to metrics.help
It covers all the common ML training metrics, and shows graphs of healthy vs unhealthy progressions. You can even paste your log output to get a quick read on your current step.
Hope you find it useful!
@yoav_farhi Love the clear graphs