Deploying PyTorch Models in Production: PyTorch Playbook

سرفصل های دوره

This course covers the important aspects of performing distributed training of PyTorch models, using the multiprocessing, data-parallel, and distributed data-parallel approaches. It also discusses which you can host PyTorch models for prediction.

1. Course Overview

1. Course Overview

2. Persisting and Loading PyTorch Models

01. Version Check

02. Module Overview

03. Prerequisites and Course Outline

04. Saving and Loading PyTorch Models

05. Building and Training a Classifier Model

06. Saving and Loading Models Using torch.save()

07. Saving Model Using the state_dict

08. Saving and Loading Checkpoints

09. Introducing ONNX

10. Exporting a Model to ONNX and Loading in Caffe2

11. Module Summary

3. Implementing Training Using Single and Multiple Processors

1. Module Overview

2. Distributed Training Options in PyTorch

3. Training Using Multiple Processes

4. Setting up a Deep Learning VM with Multiple GPUs

5. Training on Multiple GPUs

6. Module Summary

4. Implementing Distributed Training on Multiple Machines

1. Module Overview

2. Distributed Training on the Cloud

3. Setting up a SageMaker Notebook Instance

4. Setting up Training and Test Data Loaders

5. Define the Training Function

6. Functions to Test and Save the Trained Model

7. Running Distributed Training Using the PyTorch Estimator

8. Module Summary

5. Deploying PyTorch Models to Production

1. Module Overview

2. Exploring Options to Deploy PyTorch Models

3. Installing Libraries and Uploading Model Parameters to a GCP Bucket

4. Creating a Flask App to Serve the PyTorch Model

5. Using the Model for Prediction

6. Installing Docker

7. Creating and Using a Clipper Cluster for Prediction

8. Deploying a Model for Prediction to a Serverless Environment

9. Summary and Further Study

189,000 تومان

افزودن به سبد خرید

خرید دانلودی فوری

در این روش نیاز به افزودن محصول به سبد خرید و تکمیل اطلاعات نیست و شما پس از وارد کردن ایمیل خود و طی کردن مراحل پرداخت لینک های دریافت محصولات را در ایمیل خود دریافت خواهید کرد.

تولید کننده: PluralSight