Which taxa are included in the computer vision suggestions?

Modified on Mon, 13 May 2024 at 08:06 AM

This has changed over time and may change before this FAQ is updated again, as we are continually working on improving the training process. But basically, here's what's needed for a species to be included in the Computer Vision model:

  • There must be a least 100 photos of the species and 60 observations of the species, and we don't choose more than 5 photos from an observation to train the model. Observations do not need to be Research Grade in order to be used in training, but observations with a matching Community ID will be prioritized.

  • Some photos that are not included in the training phase are used to test and validate the model. These must have a Community ID.

Because of this, not every species with at least 100 photos and 60 observations will meet the requirements to be included in a training run. It's dependent on how many photos there are per observation, and whether the randomly chosen group of observations meets the requirements.

If no species within a broader taxon like genus or family meets the requirements, we may train the model on that genus or family, based on those photos.

We try to release a new version of the model once a month using transfer learning (see more here), and write a post about it in the iNaturalist Blog.

Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select atleast one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article