Frequently Asked Questions¶
Can I run this in a browser?¶
Browsers are not yet supported; you must install FiftyOne on each machine from which you want to use the library or the App.
However, check out the environments guide for best practices on using FiftyOne in common local, remote, and cloud environments.
Can I access data stored on a remote server?¶
Can I access data stored in the cloud?¶
Yes! The recommended best practice is to mount the cloud bucket to a cloud compute instance in your cloud environment and then use the remote server workflow to work with the data.
Check out the environments guide for instructions for working in AWS, GCP, and Azure.
What label types are supported?¶
FiftyOne provides support for all of the following label types for both image and video datasets:
Check out this guide for simple recipes to load labels in each of these formats.
What image file types are supported?¶
In general, FiftyOne supports all image types supported by Chromium, which includes standard image types like JPEG, PNG, TIFF, and BMP.
What video file types are supported?¶
Core methods that process videos can generally handle any codec supported by ffmpeg.
The App can play any video codec that is supported by HTML5 video on Chromium, including MP4 (H.264), WebM, and Ogg.
If you try to view a video with an unsupported codec in the App, you will be
prompted to use the
utility method to reencode the source video so it is viewable in the App.
What operating systems does FiftyOne support?¶
FiftyOne is guaranteed to support the latest versions of MacOS, Windows, and popular Linux distributions. FiftyOne will generally also support any version of these popular operating systems from the past few years.
We also provide custom install instructions to use FiftyOne on old-but-popular setups like Ubuntu 16.04 and Debian 9.
Are the Brain methods open source?¶
Check out the Brain documentation for detailed instructions on using the various Brain methods.
Can I launch multiple App instances on a machine?¶
Yes! Simply specify a different
port for each App instance that you create.
Can I connect multiple App instances to the same dataset?¶
Yes, multiple App instances can be connected to the same
Dataset via remote
Keep in mind that all users must have ssh access to the system from which the remote session(s) are launched in order to connect to them.
In addition, and all users must have FiftyOne installed on their local machines in order to launch an App instance.
You can achieve multiple connections in two ways:
Option 1: Same dataset, multiple sessions
The typical way to connect multiple App instances to the same dataset is to
create a separate remote session instance on the machine that houses the
Dataset of interest for each local App instance that you want to create.
See this FAQ for instructions on
Option 2: Same dataset, same session
Another option is to connect multiple App instances to a single remote session.
Then one or more users can use the CLI on their local machine to connect to the remote session::
# On local machine(s) # If a custom port was used, append --port XXXX fiftyone app connect --destination <username>@<remote-ip-address>
Can I connect to multiple remote sessions?¶
Yes, you can launch multiple instances of the App locally, each connected to a different remote session.
The key here is to specify a different local port for each App instance that you create.
Suppose you are connecting to multiple remote
Session instances that were
created on different remote systems (e.g., an EC2 instance and a remote server
that you own), using commands similar to:
On your local machine, you can launch App instances to
connect to the remote sessions on each
machine by specifying a different
--local-port for each App instance to use:
# Connect to first remote session fiftyone app connect --destination <username1>@<remote-ip-address1> --local-port XXXX
# Connect to second remote session fiftyone app connect --destination <username2>@<remote-ip-address2> --local-port YYYY
YYYY are any open ports on your machine.
You can also serve multiple remote sessions from the same machine.
Can I serve multiple remote sessions from a machine?¶
Yes, you can create multiple remote sessions on the same remote machine by
specifying different ports for each
Session that you create:
On your local machine(s), you can launch App instances to connect to the remote sessions that you created by specifying the corresponding remote ports that you used:
# On a local machine # Connect to first remote session fiftyone app connect \ --destination <username>@<remote-ip-address> \ --port XXXX --local-port WWWW
# On a local machine # Connect to second remote session fiftyone app connect \ --destination <username>@<remote-ip-address> \ --port YYYY --local-port ZZZZ
ZZZZ are any 4 digit ports on your local machine(s).