Timing in Deep Learning
Soumith discusses the critical role of timing in the evolution of deep learning frameworks, highlighting how TensorFlow's emergence in 2015 set a high standard for engineering quality. He reflects on the challenges of developing a fast, user-friendly package in Python, emphasizing the strategic decisions made to enhance performance. The conversation also touches on the opportunities for PyTorch in a rapidly changing landscape influenced by market dynamics.In this clip
From this podcast
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Pytorch: Fast Differentiable Dynamic Graphs in Python with Soumith Chintala - #70
Related Questions