Hyperparameter Optimization
Yi discusses the tedious nature of machine learning workflows and the importance of automation in hyperparameter tuning. He highlights a solution that utilizes randomized and grid searches, as well as Bayesian optimization through a service called Wet Lab. The conversation also touches on the significance of TensorFlow graphs as core artifacts in the system, emphasizing the need for users to manage their versioning effectively.In this clip
From this podcast
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Productizing ML at Scale at Twitter with Yi Zhaung - TWIML Talk #271
Related Questions