Skip to content
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

Jump to bottom

Fix ResNet's in_channels calculations #220

Merged
merged 2 commits into from
Nov 16, 2023
Merged

Fix ResNet's in_channels calculations #220

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f025183
fix
weihua916 Nov 16, 2023
ae8e08c
Merge branch 'master' into fix_resnet_dim
weihua916 Nov 16, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
22 changes: 13 additions & 9 deletions torch_frame/nn/models/resnet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@ def __init__(
self,
in_channels: int,
out_channels: int,
normalization: Optional[str] = 'layernorm',
normalization: Optional[str] = "layernorm",
dropout_prob: float = 0.0,
):
super().__init__()
Expand All @@ -47,10 +47,10 @@ def __init__(
self.relu = ReLU()
self.dropout = Dropout(dropout_prob)

if normalization == 'batchnorm':
if normalization == "batchnorm":
self.norm1 = BatchNorm1d(out_channels)
self.norm2 = BatchNorm1d(out_channels)
elif normalization == 'layernorm':
elif normalization == "layernorm":
self.norm1 = LayerNorm(out_channels)
self.norm2 = LayerNorm(out_channels)
else:
Expand Down Expand Up @@ -134,7 +134,7 @@ def __init__(
col_names_dict: Dict[torch_frame.stype, List[str]],
stype_encoder_dict: Optional[Dict[torch_frame.stype,
StypeEncoder]] = None,
normalization: Optional[str] = 'layernorm',
normalization: Optional[str] = "layernorm",
dropout_prob: float = 0.2,
):
super().__init__()
Expand All @@ -155,12 +155,16 @@ def __init__(
stype_encoder_dict=stype_encoder_dict,
)

in_channels = channels * (len(col_stats) - 1)
num_cols = sum(
[len(col_names) for col_names in col_names_dict.values()])
in_channels = channels * num_cols
self.backbone = Sequential(*[
FCResidualBlock(in_channels if i == 0 else channels, channels,
normalization=normalization,
dropout_prob=dropout_prob)
for i in range(num_layers)
FCResidualBlock(
in_channels if i == 0 else channels,
channels,
normalization=normalization,
dropout_prob=dropout_prob,
) for i in range(num_layers)
])

self.decoder = Sequential(
Expand Down