We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
add assert, 0.0.24
simplify even more by making VAE required and inferring the number of… … image tokens and image sequence length off the VAE
fix generation and further cleanup by moving generate method into Dal… …le class
make sure generation is done in eval mode, and sets it back to origin… …al state
use state of the art vision transformer in CLIP - 0.0.18
make sure last token can still predict eos
allow for transparent images