Skip to content

[Contrib][TRT] Fix Conv2D construction when channels attribute is not available. #6805

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
tqchen merged 1 commit into from
Oct 31, 2020
Merged

[Contrib][TRT] Fix Conv2D construction when channels attribute is not available. #6805

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
2 changes: 1 addition & 1 deletion src/runtime/contrib/tensorrt/tensorrt_ops.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -242,7 +242,7 @@ class Conv2DOpConverter : public TensorRTOpConverter {
auto str_dilation = params->node.GetAttr<std::vector<std::string>>("dilation");
auto str_padding = params->node.GetAttr<std::vector<std::string>>("padding");
int groups = std::stoi(params->node.GetAttr<std::vector<std::string>>("groups")[0]);
int channels = std::stoi(params->node.GetAttr<std::vector<std::string>>("channels")[0]);
int channels = weight_shape[0];
// TRT conv2d op doesn't support asymmetric padding before 5.1, so we
// workaround by adding a padding layer before the pooling op.
nvinfer1::DimsHW prepadding, postpadding;
Expand Down
1 change: 0 additions & 1 deletion tests/python/contrib/test_tensorrt.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -251,7 +251,6 @@ def get_graph(
out = relay.nn.conv2d(
x,
kernel,
channels=k_shape[0],
kernel_size=k_shape[2:4],
groups=groups,
padding=padding,
Expand Down