Explore the journey from NumPy to PyTorch in this insightful 26-minute talk from PyCon US. Delve into the concept of "NumPy-compatibility" in PyTorch, understanding its significance, challenges, and occasional need for divergence. Learn why the PyTorch community values this compatibility and how it impacts the broader data science ecosystem. Gain valuable insights into the intricacies of API design and the balance between maintaining familiarity and introducing innovation. Perfect for data scientists, machine learning engineers, and anyone interested in the evolution of scientific computing libraries.
From NumPy to PyTorch - A Story of API Compatibility
Overview
Syllabus
TALK / Randall Hunt, Mike Ruberry / From NumPy to PyTorch, A Story of API Compatibility
Taught by
PyCon US
Related Courses
-
Python Numpy: Machine Learning & Data Science Course
-
Small Big Data - Using NumPy and Pandas When Your Data Doesn't Fit in Memory
-
Building NumPy Arrays from CSV Files, Faster than Pandas
-
GPU-Accelerated Data Analytics in Python
-
Skynet 101 - How to Keep Your Machine Learning Code From Getting Away From You
-
Python in High Energy Physics
