Deep Learning Challenges
Aravind recounts the challenges of early deep learning, working with C++ and Cuda before Python frameworks like Tensorflow. Debugging was a nightmare, requiring rewriting layers in C++ and reinstalling Cuda drivers. The evolution to Python-based frameworks like Pytorch brought relief from the painstaking process.In this clip
From this podcast
Invest Like the Best
Aravind Srinivas - Building An Answer Engine - [Invest Like the Best, EP.363]
Related Questions