Scribd Logo
Upload

Welcome to Scribd!

  • Lec 3

    Uploaded by

    张竞择
    9 views7 pages
    This document provides an overview and introduction to PyTorch for a deep reinforcement learning course. It explains that the goal of the course is to train agents to perform useful tasks. It then introduces PyTorch as a Python library for defining neural networks, computing gradients automatically, and other functions like working with datasets and optimizers. The document explains that PyTorch allows you to define a neural network and it will automatically compute the gradients to train the model, focusing on the key aspects of defining and training neural networks. It concludes by providing a link to a PyTorch tutorial on Colab.

    Original Description:

    Original Title

    lec-3

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    This document provides an overview and introduction to PyTorch for a deep reinforcement learning course. It explains that the goal of the course is to train agents to perform useful tasks. It then introduces PyTorch as a Python library for defining neural networks, computing gradients automatically, and other functions like working with datasets and optimizers. The document explains that PyTorch allows you to define a neural network and it will automatically compute the gradients to train the model, focusing on the key aspects of defining and training neural networks. It concludes by providing a link to a PyTorch tutorial on Colab.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    9 views7 pages

    Lec 3

    Uploaded by

    张竞择
    This document provides an overview and introduction to PyTorch for a deep reinforcement learning course. It explains that the goal of the course is to train agents to perform useful tasks. It then introduces PyTorch as a Python library for defining neural networks, computing gradients automatically, and other functions like working with datasets and optimizers. The document explains that PyTorch allows you to define a neural network and it will automatically compute the gradients to train the model, focusing on the key aspects of defining and training neural networks. It concludes by providing a link to a PyTorch tutorial on Colab.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    of 7

    PyTorch and Neural Nets

    CS285 Deep RL
    Instructor: Marwa Abdulhai

    [Adapted from Kevin Li’s CS285 Fa21 Slides]


    Goal of this course
    Train an agent to perform useful tasks

    𝜋! (𝑎|𝑠) )"#$ = 𝑓! (𝑠" , 𝑎" )


    ∆
    𝑄! (𝑠, 𝑎)
    train the model

    data agent

    collect data
    Goal of this course
    Train an agent to perform useful tasks

    𝜋! (𝑎|𝑠) )"#$ = 𝑓! (𝑠" , 𝑎" )


    ∆
    𝑄! (𝑠, 𝑎)
    train the
    train model
    the model

    focus for today’s lecture!


    data agent

    collect data
    How do train a model?
    𝜃 ∗ = arg min" ∑ #,% ∈' ℒ( 𝑓" 𝑥 , 𝑦)

    dataset neural network


    gradient
    descent loss

    PyTorch does all of these!


    What is PyTorch?

    Python library for:


    • Defining neural networks
    • Automating computing gradients
    • And more! (datasets, optimizers,
    GPUs, etc.)
    How does PyTorch work?

    You define: ℎ! = 𝜎 𝑊! 𝑥 ℎ" = 𝜎 𝑊" ℎ! 𝑦 = 𝜎 𝑊# ℎ"


    $% $% $'" $'! $% $% $'" 𝜕𝑦
    PyTorch computes: = =
    $&! $'" $'! $&! $&" $'" $&! 𝜕𝑊#

    [picture from Stanford’s CS231n]


    PyTorch Tutorial (Colab)
    https://colab.research.google.com/drive/1XQu1mUbGtvkQY-D7_YCOZlRzSnjp4u9f?usp=sharing

    https://bit.ly/3CM6lcf

    You might also like