To download original ImageNet and Pascal VOC datasets, follow the instructions below for each dataset type. These datasets are considerably big in size. If you want to save time when loading them into the DL Workbench, you can cut an original dataset.
To learn more about dataset types supported by the DL Workbench and their structure, refer to Dataset Types.
To download images from ImageNet, you need to have an account and agree to their Terms of Access. Follow the steps below:
Download a script to cut datasets. In a Python* console, run the following command after specifying the parameters:
This command runs the script with the following arguments:
Parameter | Explanation |
---|---|
--source_archive_dir=C:\Users\Work\imagenet.zip |
Full path to a downloaded archive |
--output_size=10 |
Number of images to be left in a smaller dataset |
--output_archive_dir=C:\Users\Work\subsets |
Full directory to the smaller dataset, excluding the name |
--dataset_type=imagenet |
Type of the source dataset |
To download test data from Pascal VOC, you need to have an account. Follow the steps below:
Download a script to cut datasets. In a Python* console, run the following command after specifying the parameters:
This command runs the script with the following arguments:
Parameter | Explanation |
---|---|
--source_archive_dir=C:\Users\Work\voc.tar.gz |
Full path to a downloaded archive |
--output_size=10 |
Number of images to be left in a smaller dataset |
--output_archive_dir=C:\Users\Work\subsets |
Full directory to the smaller dataset, excluding the name |
--dataset_type=voc |
Type of the source dataset |