00:00 / 00:00
Sleep timer ended
⏪ 10s
⏩ 10s
2.8K views
3 weeks ago
Sebastian Russo, Ivan Llanos, and Sahil Dua from Google DeepMind introduce EmbeddingGemma 2, a lightweight open model that maps text, images, video, and audio i…
...more
17 Comments
if only I had a use case ! I like this idea
isn't this old though...?
You should hire actors, I don't need to see the actual face of the product leads.
Great. Thanks for sharing this
Cool. <br>The naming though - why not make it part of the Gemma 4 family.
And google cooked, again...
Imagine a world without Google... yeah, that will be hell.
Cool
this is nice
Wonder how well it will handle FPGAs then
❤❤❤
Awesome! I can't wait to do the test for this model, and use it for regonize many image and more mov on phone
Mapping text, audio, and visual data into a single unified embedding space under 1B parameters is a huge win for local-first architectures. Combining native multimodal alignment with Matryoshka representation down to 128 dimensions makes on-device vector search genuinely practical for resource-constrained edge devices. Excellent release!
damn so cool
love from open-source community <3
Noice
Lik anna
Up next
From this channel
