Skip to content

Fix value_dim in TransformerDecoder's cross-attn layer #667

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

Merged
merged 2 commits into from
Jan 18, 2023
Merged

Fix value_dim in TransformerDecoder's cross-attn layer #667

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e50c17b
Fix in TransformerDecoder's cross-attn layer
abheesht17 Jan 16, 2023
9a9449c
Revert UT
abheesht17 Jan 18, 2023
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
2 changes: 1 addition & 1 deletion keras_nlp/layers/transformer_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,7 +165,7 @@ def _build(self, input_shape, has_cross_attention):
self._cross_attention_layer = keras.layers.MultiHeadAttention(
num_heads=self.num_heads,
key_dim=head_dim,
value_dim=hidden_dim,
value_dim=head_dim,
dropout=self.dropout,
kernel_initializer=clone_initializer(self.kernel_initializer),
bias_initializer=clone_initializer(self.bias_initializer),
Expand Down
4 changes: 3 additions & 1 deletion keras_nlp/layers/transformer_decoder_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,9 @@ class MyModel(keras.Model):
def __init__(self):
super().__init__()
self._decoder = transformer_decoder.TransformerDecoder(
intermediate_dim=4, num_heads=2
intermediate_dim=4,
num_heads=2,
has_cross_attention=True,
Copy link
Member

@mattdangerw mattdangerw Jan 17, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, why do we need this actually? Won't this line at the start of call has_encoder_sequence = encoder_sequence is not None, mean the layer will be built with cross attention as soon as the decoder is called on two inputs?

Copy link
Collaborator Author

@abheesht17 abheesht17 Jan 18, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry 🤦🏼 , not needed. Changing it back

)
self._dense = keras.layers.Dense(1, activation="sigmoid")

Expand Down