Skip to content

Fix MultiwayEmbedding class name typo #125

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Hashim Khan (Hashim1999164) wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix MultiwayEmbedding class name typo #125

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 5 additions & 2 deletions torchscale/component/multiway_network.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,11 +45,14 @@ def forward(self, x, **kwargs):
return torch.cat([y1, y2], dim=self.dim)


class MutliwayEmbedding(MultiwayNetwork):
class MultiwayEmbedding(MultiwayNetwork):
def __init__(self, modules, dim=1):
super(MultiwayNetwork, self).__init__()
self.dim = dim
assert len(modules) == 2
self.A = modules[0]
self.B = modules[1]
self.split_position = -1
self.split_position = -1

# Backwards-compatible alias for the previous misspelling.
MutliwayEmbedding = MultiwayEmbedding
4 changes: 2 additions & 2 deletions torchscale/model/BEiT3.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@
TextEmbedding,
VisionEmbedding,
)
from torchscale.component.multiway_network import MutliwayEmbedding
from torchscale.component.multiway_network import MultiwayEmbedding


class BEiT3(nn.Module):
Expand All @@ -30,7 +30,7 @@ def __init__(self, args, **kwargs):
prepend_cls_token=True,
)
# being consistent with Fairseq, which starts from 2 for position embedding
embed_positions = MutliwayEmbedding(
embed_positions = MultiwayEmbedding(
modules=[
PositionalEmbedding(self.vision_embed.num_position_embeddings() + 2, args.encoder_embed_dim),
PositionalEmbedding(args.max_source_positions, args.encoder_embed_dim),
Expand Down