X-posting as of special Commons interest.
On 27 September 2016 at 01:57, Pine W wiki.pine@gmail.com wrote:
Perhaps of interest: "...We’re making the latest version of our image captioning system available as an open source model in TensorFlow." https://research.googleblog.com/2016/09/show-and-tell-image-captioning-open....
Pine
I wonder if anyone would be prepared to set this up in Wikimedia-Labs as a callable tool to test out? This would be great for, say, pre-filtering Flickr uploads to help avoid copyright issues or limit photographs to those of particular educational value. It's a set of Python modules, but I'm unsure of how complex the dependencies would be.
I'm currently uploading someone's 8,000 photograph stream, and it would great to limit it to just the ones of birds, plants and landscapes, leaving out the shots of people on holiday or family photos.
Fae