Skip to content

Commit

Permalink
[docs] Update ML article
Browse files Browse the repository at this point in the history
  • Loading branch information
vishnukvmd committed Sep 28, 2024
1 parent 05763a5 commit f255ded
Showing 1 changed file with 19 additions and 23 deletions.
42 changes: 19 additions & 23 deletions docs/docs/photos/features/machine-learning.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,45 +7,41 @@ description:

# Machine learning

> [!NOTE]
>
> This document describes a beta feature that will be present in an upcoming
> release.
Ente supports on-device machine learning. This allows you to use the latest
advances in AI in a privacy preserving manner.

- You can search for your photos by the **faces** of the people in them. Ente
- You can search for your photos by the **Faces** of the people in them. Ente
will show you all the faces in a photo, and will also try to group similar
faces together to create clusters of people so that you can give them names,
and quickly find all photos with a given person in them.

- You can search for your photos by typing natural language descriptions of
them. For example, you can search for "night", "by the seaside", or "the red
motorcycle next to a fountain". Within the app, this ability is sometimes
referred to as **magic search**.
motorcycle next to a fountain". Within the app, this ability is referred to
as **Magic search**.

- We will build on this foundation to add more forms of advanced search.
You can enable face recognition and magic search in the app's preferences on
either the mobile app or the desktop app.

You can enable face and magic search in the app's preferences on either the
mobile app or the desktop app.
On mobile, this is available under `General > Advanced > Machine learning`.

If you have a big library, we recommend enabling this on the desktop app first,
because it can index your existing photos faster (The app needs to download your
originals to index them which can happen faster over WiFi, and indexing is also
faster on your computer as compared to your mobile device).
On desktop, this is available under `Preferences > Machine learning`.

---

Once your existing photos have been indexed, then you can use either. The mobile
app is fast enough to easily and seamlessly index the new photos that you take.
The app needs to download your original photos to index them. This is faster
over WiFi. Indexing is also faster on your computer as compared to your mobile
device.

> [!TIP]
>
> Even for the initial indexing, you don't necessarily need the desktop app, it
> just will be a bit faster.
> If you have a large library on Ente, we recommend enabling this feature on the
> desktop app first, because it can index your existing photos faster. Once your
> existing photos have been indexed, then you can use either. The mobile app is
> fast enough to index new photos as they are being backed up.
The indexes are synced across all your devices automatically using the same
end-to-end encypted security that we use for syncing your photos.
end-to-end encrypted security that we use for syncing your photos.

Note that the desktop app does not currently support viewing and modifying the
automatically generated face groupings, that is only supported by the mobile
app.
Note that the desktop app does not currently support modifying the face
groupings, that is only supported by the mobile app.

0 comments on commit f255ded

Please sign in to comment.