github

rasbt / scipy2023-deeplearning

  • четверг, 13 июля 2023 г. в 00:00:06
https://github.com/rasbt/scipy2023-deeplearning



SciPy 2023 Workshop

Modern Deep Learning with PyTorch

At SciPy in Austin, Texas

Mon 10 July 2023, 1:30–5:30 pm (CDT, Chicago, local time), Classroom 202

Abstract

We will kick off this tutorial with an introduction to deep learning and highlight its primary strengths and use cases compared to traditional machine learning. In recent years, PyTorch has emerged as the most widely used deep learning library for research. However, a lot has changed regarding how we train neural networks these days. After getting a firm grasp of the PyTorch API, you will learn how to train deep neural networks using various multi-GPU training paradigms. We will also fine-tune large language models (transformers)!

Material & Preparation

The workshop material will be posted on the weekend before the event. To prepare for the workshop, there are only 3 small action items

  1. (Optional) You may find the Python Setup Guide (./00-1_python-setup-guide) helpful, which mainly describes how I set up Python on my computer(s).
  2. Please go through Python Library Installation (./00-2_python-libraries-for-workshop) guide to ensure you have all the required libraries installed prior to the workshop.
  3. I recommend downloading this repository before the event so you can access the materials offline in case of a slow internet connection during the workshop.

Looking forward to seeing you there!

PS: If you have any questions, please feel free to reach out via the Discussion page here on GitHub.

Schedule and Slides

  1. Introduction to Deep Learning (1:30 - 2:00 pm) [Slides]
  2. Understanding the PyTorch API (2:00 - 2:30 pm) [Slides]
  3. Training Deep Neural Networks (2:30 - 3:00 pm) [Slides]

10 Min Break

  1. Accelerating PyTorch Model Training (3:10 - 3:45 pm) [Slides]
  2. Organizing PyTorch Code (3:45 - 4:15 pm) [Slides]
  3. More Tips and Techniques (4:15 - 4:45 pm) [Slides]

10 Min Break

  1. Finetuning LLMs (4:55 - 5:25 pm) [Slides]
  2. Wrap Up & Questions (5:25 - 5:30 pm) [Slides]