Is it possible to access the embeddings directly rather than only predictions so that they can be applied to other downstream tasks ?
hi, I am also interesting in model’s embedding, I want to use it to get phage sequence’s embedding, because it’s too long to handle with for me.
Same interesting here too!
Hi all,
Sorry we don’t currently provide a means to request the underlying trunk embeddings.
Hi! I wanted to follow up on this to see if accessing the trunk embeddings is or will soon be an available feature?
I also wanted to say I really appreciate all of your documentation and efforts to make your science open source!! It is very inspiring to see!
Thanks!
Our model code and weights are now available at https://github.com/google-deepmind/alphagenome_research, so you should be able to run the model locally and access the trunk embeddings.