I found this and this on embeddings from Google.
Here is the live TensorBoard Embedding Projector without the need to install tensorflow.
You can use your own dataset or try the datasets available in the visualizer as an example.
I found this and this on embeddings from Google.
Here is the live TensorBoard Embedding Projector without the need to install tensorflow.
You can use your own dataset or try the datasets available in the visualizer as an example.
That’s pretty cool, been looking for something like this for a while. But just to be sure I understand:
Have you tried to do it ?
I have the intention to do this one day, but havn’t done it yet
Converting pytorch embeddings into TSV format needs some work. People have done the reverse (reading TSV files into pytorch):