datasets
datasets copied to clipboard
Add `tfds list` CLI command
Addition of community dataset will sharply increase the no of available datasets. So it will become difficult to browse through repos or catalog.
So, it would great if we can add a new CLI command to list all the datasets in present in TFDS. This will allows users to easily search a dataset using tools like grep.
Possible Usage
tfds list # List all datasets
tfds list --type audio # List all audio dataset in TFDS
tfds list --namespace huggingface # List all huggingface community dataset
tfds list --search sun # Lists datasets like sun397, lsun (alternative is to use grep)
Implementation details
• Use/modify the tfds.list_builder
API.
• Add a new file list.py
in scripts/cli
folder
Hi, I have started work on this
Since @jatin-code777 has already started working on --type
and --exclude_community
part, I will be sending a PR for the
--search
and --namespace
flags soon.
I am working on this issue. Will you please assign it to me?
Hey @Srikeshram, @jatin-code777 is already working on the issue and taking it up would only lead to work duplication. You can take up other bugs and issues which are still unresolved in TFDS. Thank you for the efforts!
Leaving here my cheers on this feature request!
It'd be a game-changer to be able to dynamically do "dataset version control" easier.