“I have been using wandb to keep track of my experiments and results for a while now. And it is easy to install and intuitive to use. I have been recommending this to my students for 2 years now!”

“Weights and Biases is one of the tools in my PhD that I cannot live without and every training code I write, always starts with calling import wandb. I only request easier offline service similar to tensorboard without requiring huge setup so can be for corporate projects where online logging is not permitted.”

“I DO NOT recommend Weights and Bias! It's impossible to uninstall. Out of curiosity, I signed up, installed its packages, and linked it with Kaggle account to try it out. It was good, but it only tracks model training progress. I much prefer MLflow's end-to-end pipeline. I decided to delete it, but W&B has its packages installed on Kaggle by default. So, by configuration, whenever I train a model, I'm prompted to login to W&B. I've posted on the W&B Discord. But, their staff never answer questions the any channel.”

“I use it for all my experiments tracking! It really revoluzionized the way I keep track of progress and allowed me to focus much more on other tasks”

“Been using W&B for years and cannot recommend it enough. Very straightforward to use, incredibly useful and pretty reasonable pricing (free even for personal and academic use afaik).”

“Positive feedback received via Google Places.”