Deep Learning with PyTorch course
Get started quickly Explained how to use it HyperAI Run the first one JupyterLab. Here we will specifically introduce how to use it HyperAI of Jupyter Clone a workspace with a public tutorial.
Deep Learning with PyTorch: A 60 Minute Blitz yes Pytorch The official learning tutorial. Introduced it Pytorch Some basic grammar. Concept and introduction of adoption cifar10 The process of training an image classification model on a dataset. It provides Google Colab The execution method also allows downloading Jupyter of .ipynb file. Openbayes It also provides a compiled and localized version. You can directly access it HyperAI Below Jupyter Used in the workspace.
Clone Public Tutorial
In login HyperAI Then select the one on the left "public resource " - "Public Tutorial" after, search "PyTorch Official Tutorial use PyTorch Implementing deep learning" .
You can see that this tutorial adopts pytorch 1.9 And used it vGPU Type of container. Click on the top right corner "clone" Clone the container as a new container.
As shown in the above figure. After cloning the container, you can choose freely "Computing power resources" and "Runtime environment" .
If the public tutorial is also bound to other publicly available models or datasets, "clone" When it comes to binding, it will also be bound by default.
After a while, you can see an open container:
click "open Jupyter working space" Ready to use JupyterLab working space.
If you see the following prompt. Please click on allow. This feature will support JupyterLab Browser reminder for completed execution.
If you don't know how to use it yet JupyterLab. You can refer to our Jupyter working space file, JupyterLab Official documents Or related Chinese translation materials.
Execute Tutorial
In the open Jupyter Double click the file in the left-hand directory of the workspace Deep Learning with PyTorch.ipynb Open this file.
Then you can browse the tutorial and learn each chapter one by one.
If you need to restart the tutorial. You can select from the top navigation bar "Kernel" - "Restart the kernel" .
If you want to execute it all at once ipynb All codes in the file can be selected "Kernel" - "Restart the service and run all code blocks" .
Download file
After a series of editing work. We need to update the ones that have already been updated .ipynb Download locally. There are multiple ways to download.
From Jupyter Download directly from the workspace
For those that are currently running Jupyter working space. You can right-click on the file you want to download. Click to download.
If you want to download the entire directory, you can right-click in the blank space of the left navigation and select "Package download folder" It will automatically package and download the current directory:
Download from the working directory tab
If Jupyter The workspace has been closed. You can access it through the page "output" Tab for downloading. You can choose from them on the right side "Download current directory" .
You can also download individual files directly through the download button on the right side of the file list.
The container in operation "output" The files under the tab will periodically synchronize data from the container. Therefore, compared to the container. The update time of its files will be later than the files in the container. It is recommended to download the file from here after the container execution is completed. To avoid obtaining incomplete file content.