rewrite the mace equivariant head to avoid a model explosion #52
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
original model:
time embedding + mace output ->
o3.Linear -> o3.BatchNorm -> o3.TensorSquare -> non linearity
The TensorSquare had a lot of weights and was probably overkill for our purpose. Its goal was to mix the time information across the different channels of MACE, but it was also computing a lot of useless stuff.
I replaced with the following:
FullyConnectedTensorProduct(time embedding, mace output) ->
o3.Linear -> o3.BatchNorm -> non linearity
the FCTP mixes time with all the components of the mace output and we can control the dimensionality of its output. No need for a TensorSquare anymore
We probably could simplify even more by allowing only the 0e channel as the output of the FCTP since the next layers should not mix different ells