Add Upsample Layer #2255
Merged
Add Upsample Layer #2255
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
fiedorowicz1
force-pushed
the
add-upsample-layer
branch
from
60a182d to
f0fda83
Compare
fiedorowicz1
force-pushed
the
add-upsample-layer
branch
from
edb7f9f to
8dc219b
Compare
benson31
approved these changes
Jun 24, 2024
Comment on lines
31
to
38
| template <typename TensorDataType, data_layout Layout, El::Device Device> | ||
| template <typename ArchiveT> | ||
| void upsample_layer<TensorDataType, Layout, Device>::serialize(ArchiveT& ar) | ||
| { | ||
| using DataTypeLayer = data_type_layer<TensorDataType>; | ||
| ar(::cereal::make_nvp("DataTypeLayer", | ||
| ::cereal::base_class<DataTypeLayer>(this))); | ||
| } |
There was a problem hiding this comment.
Should this save more of the state? At least the upscale mode and scaling factors?
src/layers/transform/upsample.cpp
Outdated
|
|
||
| /// Pooling forward propagation with im2col | ||
| template <typename TensorDataType, data_layout Layout, El::Device Dev> | ||
| void upsample_layer<TensorDataType, Layout, Dev>::fp_compute_im2col() |
There was a problem hiding this comment.
You might consider just deleting this and the corresponding bp_ function to just remove unnecessary code. The call-sites are commented out.
src/layers/transform/upsample.cpp
Outdated
|
|
||
| /// Pooling forward propagation with im2col | ||
| template <typename TensorDataType, data_layout Layout, El::Device Dev> | ||
| void upsample_layer<TensorDataType, Layout, Dev>::bp_compute_im2col() |
Co-authored-by: Tom Benson <[email protected]>
benson31
reviewed
Jun 24, 2024
| #include "lbann/utils/dnn_lib/upsample.hpp" | ||
| #endif // LBANN_HAS_DNN_LIB | ||
| #include "lbann/utils/exception.hpp" | ||
| #include "lbann/utils/im2col.hpp" |
There was a problem hiding this comment.
Suggested change
| #include "lbann/utils/im2col.hpp" |
Comment on lines
270
to
275
| /// Pooling forward propagation with im2col | ||
| void fp_compute_im2col(); | ||
|
|
||
| /// Pooling forward propagation with im2col | ||
| void bp_compute_im2col(); | ||
|
|
There was a problem hiding this comment.
Suggested change
| /// Pooling forward propagation with im2col | |
| void fp_compute_im2col(); | |
| /// Pooling forward propagation with im2col | |
| void bp_compute_im2col(); |
Comment on lines
264
to
267
| /// Pooling forward propagation with DNN library | ||
| void fp_compute_dnn(); | ||
|
|
||
| /// Pooling backward propagation with DNN library |
There was a problem hiding this comment.
Suggested change
| /// Pooling forward propagation with DNN library | |
| void fp_compute_dnn(); | |
| /// Pooling backward propagation with DNN library | |
| /// Upscaling forward propagation with DNN library | |
| void fp_compute_dnn(); | |
| /// Upscaling backward propagation with DNN library |
tbennun
requested changes
Jun 25, 2024
Co-authored-by: Tal Ben-Nun <[email protected]>
tbennun
approved these changes
Jun 25, 2024
bvanessen
approved these changes
Jun 28, 2024
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.
Adds a distconv supported upsampling layer. Currently, this layer can only perform nearest neighbor upsampling and only works on NVIDIA machines.