You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was trying to identify the projector component in Pi0 (the module that projects visual features from the vision encoder to the input space of the LLM). Just wanted to ask to make sure:
Is the projector applied in the following lines? (scr/openpi/models/siglip.py, lines 284-288):
if self.num_classes:
kw = {"kernel_init": nn.initializers.zeros} if self.head_zeroinit else {}
head = nn.Dense(self.num_classes, dtype=self.dtype_mm, name="head", **kw)
x_2d = out["logits_2d"] = head(x_2d)
x = out["logits"] = head(x)
that is, the main output of the vision encoder is already passed through the projector?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
I was trying to identify the projector component in Pi0 (the module that projects visual features from the vision encoder to the input space of the LLM). Just wanted to ask to make sure:
Is the projector applied in the following lines? (scr/openpi/models/siglip.py, lines 284-288):
that is, the main output of the vision encoder is already passed through the projector?
Beta Was this translation helpful? Give feedback.
All reactions