Platform For AI:Submit a single-node PyTorch transfer learning job

Step 1: Get the data

The sample data is available at a public URL (Download the data) and requires no preparation.

Step 2: Get the training code

The training code is available at a public URL (Download the training code). No development required.

Step 3: Create a job

1. Go to the Create Job page.

   1. Log on to the PAI console. In the top navigation bar, select a region. In the upper-right corner, select a workspace, and click Go to DLC.

   2. On the Deep Learning Containers (DLC) page, click Create Task.

2. On the Create Task page, configure the following parameters. Use default values for other settings.

   

   Parameter		Description
   Basic Information	Task Name	Enter the job name, such as torch-sample.
   Environment Information	Node Image	Click Alibaba Cloud Image and select a PyTorch image.
   	Datasets	Mount a custom dataset to save training results. This example uses an OSS dataset. Click Custom Dataset and configure the following parameters: Custom Dataset: Select an existing OSS dataset. Create and manage datasets. Mount Path: Set the path to /mnt/data/.
   	Start Command	Enter the following command to download the data and code, run the training job, verify the output, and copy results to the mounted directory. wget https://pai-public-data.oss-cn-beijing.aliyuncs.com/hol-pytorch-transfer-cv/data.tar.gz && tar -xf ./data.tar.gz && mv ./hymenoptera_data/ ./input && mkdir output && wget https://pai-public-data.oss-cn-beijing.aliyuncs.com/hol-pytorch-transfer-cv/main.py && python main.py -i ./input -o ./output && ls ./output && cp -r ./output /mnt/data
   Resource Information	Resource Source	Select Public Resources.
   	Framework	Select PyTorch.
   	Task Resources	Nodes: Set to 1. Instance Type: Click and select an instance type, such as GPU > ecs.gn6e-c12g1.3xlarge. If this instance type is unavailable in your region, switch to a supported region. Deep Learning Containers (DLC) lists pay-as-you-go regions.

3. Click OK.

   You are automatically redirected to the Deep Learning Containers (DLC) page.

Step 4: View job details and logs

1. On the Deep Learning Containers (DLC) page, click the job name.

2. On the job overview page, view Basic Information and Resource Information.

3. Scroll down to the Instance section. In the Actions column, click Log.

4. Output files are stored in the mounted dataset's file system. The following figure shows example output in an OSS bucket. Your results may vary.