Addressing Memory Overflow Issues with Large Datasets

Подробнее о видео

Squeezing a large dataset from a Parquet or CSV file into a pandas DataFrame is no fun when your data is too large for memory. This video presents three solutions to get your data to fit into a pandas DataFrame.

💡Hint: The third solution is the best if you are on Google Colab, Kaggle, or simply using a GPU.

00:00 - Intro
00:30 - Solution 1 - Swap Space
02:04 - Solution 2 - Sampling
06:06 - Solution 3 - Unified Virtual Sampling
7:43 - nv_dashboard
8:19 - UVM - Getting Started
10:19 - Pros / Cons of Solutions

📝 Learn more about cuDF UVM:

➡️ Join the NVIDIA Developer Program:

➡️ Read and subscribe to the NVIDIA Technical Blog:

#pandas
#datascience
#gpu
#python
#nvidia
#cudf