ever wondered how the text search on your phone image gallery works?
@AIatMeta released MetaCLIP2, and we've added it to @huggingface transformers 🔥
it's a multilingual model that can understand image + text!
find the notebook for text-to-image search on the next ⤵️
@aryanvs_ lots of in-game env-physics simulations, and heavy feedback for pretraining i guess
like craft literaly millions sims in unreal, houdini, etc, and start pretrain there? then go towards actual-real-footage later
atleast in this one it looks really gamey, but with a real 'look'
We first encode the source sample to create a codebook of latents as our timbre reference.
Then after encoding our target sample into a sequence of latents, we reconstruct that sequence instead using the closest elements from our codebook 👾
Show don't tell...
Magenta RealTime now supports audio prompting in addition to text prompting, so you can now use clips of audio as latent anchors to steer generation.
https://t.co/AFtzT92Zvi