Add const qualifier to isNullTensor()#446
Merged
kevinch-nv merged 1 commit intoonnx:masterfrom Apr 23, 2020
Merged
Conversation
kevinch-nv
added a commit
that referenced
this pull request
Jul 1, 2020
* Updating linear resize dimensions check (#362) * Updates global pooling functions to work correctly with dynamic shapes (#365) * Support empty initializers for optional inputs (#366) * Add support for empty initializers for optional inputs * Alphabetize importPluginFactory * Support ceiling mode padding for dynamic inputs (#368) * Register empty Constant node outputs to support empty weights (#369) * Update myelin library name on Windows (#371) * Update logic to import ONNX initializers (#375) * Adding more type checks (#380) * Add type check for gather and shapedweights attribute imports (#384) * Throw warning if seed input is provided for randomuniform nodes (#386) * Update spacetodepth importer to support fulldims and dynamic shapes (#392) * Add check to avoid console spam of warnings (#402) * fix some build warnings/errors on Windows VS2019 (#403) * remove c++11/14 non-compliant constexpr lambdas * fix build warnings on VS2019 * disable shape input tensor * Revert "disable shape input tensor" This reverts commit 9a49e03. * Support opset11 padding (#408) * Fix loop importer scan output calculation (#412) * Fix typo in operators.md supported onnx operators (#399) There are two overlapping RNN operators, one supporting and the other not. Since onnx supports RNN, the one with supported N should be removed. Signed-off-by: juhyung <sonju0427@gmail.com> * Added optimization only mode which runs optimization passes on the model without converting it to tensorrt. (#420) * New command line options. * Updated documentation. * Currently requires linking against onnx project. * Support opset8 scan (#433) * Fix deconv importer and remote instancenormalization epsilon clamp value (#434) * Fix deconv importer and remote instancenormalization epsilon clamp value * Remove dilations * Add check for shape tensor outputs (#437) * Fix slice caclulation for -INT_MAX (#438) * Support boolean weight conversion to tensors (#439) * Fix node output accesser for older versions of protobuf (#441) * Add const qualifier to isNullTensor() (#446) * Support negative slicing across an entire axis (#453) * Keep track of Loop tensor mappings (#454) * Fix fp16 weight import (#484) * Fix GEMM import assertion (#485) Co-authored-by: pranavm-nvidia <49246958+pranavm-nvidia@users.noreply.github.com> Co-authored-by: George Wu <jywu@microsoft.com> Co-authored-by: JuHyung Son <sonju0427@gmail.com> Co-authored-by: Dennis Sandler <sandler.denis@gmail.com>
tdp2110
added a commit
to tdp2110/onnx-tensorrt
that referenced
this pull request
Jul 9, 2020
Suppose we have a network with (not all distinct) layer names layer layer_1 layer When ImporterContext sees "layer", it sees it's not in mLayerNameCounts, and sets mLayerNameCounts["layer"] = 1 and adds a TRT layer with name "layer". It then sees "layer_1", concludes it's not in mLayerNameCounts, so it sets mLayerNameCounts["layer_1"] = 1 and adds a TRT layer with name "layer_1". NOW when it sees "layer", it sees that mLayerNameCounts["layer"] == 1, so we produce a "uniqueName" of "layer" + "_" + std::to_string(mLayerNameCounts["layer"] ), ie "layer_1", which is a name conflict for the TRT net. This change keeps track of all inserted names in a set and in the case of duplicates, suffix-appended modifications of the duplicated name by ever increasing integers until a name appears which has not been used.
tdp2110
added a commit
to tdp2110/onnx-tensorrt
that referenced
this pull request
Jul 9, 2020
Suppose we have a network with (not all distinct) layer names layer layer_1 layer When ImporterContext sees "layer", it sees it's not in mLayerNameCounts, and sets mLayerNameCounts["layer"] = 1 and adds a TRT layer with name "layer". It then sees "layer_1", concludes it's not in mLayerNameCounts, so it sets mLayerNameCounts["layer_1"] = 1 and adds a TRT layer with name "layer_1". NOW when it sees "layer", it sees that mLayerNameCounts["layer"] == 1, so we produce a "uniqueName" of "layer" + "_" + std::to_string(mLayerNameCounts["layer"] ), ie "layer_1", which is a name conflict for the TRT net. This change keeps track of all inserted names in a set and in the case of duplicates, tries suffix-appended modifications of the duplicated name by ever increasing integers until a name appears which has not been used.
kevinch-nv
pushed a commit
that referenced
this pull request
Jul 10, 2020
Suppose we have a network with (not all distinct) layer names layer layer_1 layer When ImporterContext sees "layer", it sees it's not in mLayerNameCounts, and sets mLayerNameCounts["layer"] = 1 and adds a TRT layer with name "layer". It then sees "layer_1", concludes it's not in mLayerNameCounts, so it sets mLayerNameCounts["layer_1"] = 1 and adds a TRT layer with name "layer_1". NOW when it sees "layer", it sees that mLayerNameCounts["layer"] == 1, so we produce a "uniqueName" of "layer" + "_" + std::to_string(mLayerNameCounts["layer"] ), ie "layer_1", which is a name conflict for the TRT net. This change keeps track of all inserted names in a set and in the case of duplicates, tries suffix-appended modifications of the duplicated name by ever increasing integers until a name appears which has not been used.
kevinch-nv
added a commit
that referenced
this pull request
Oct 20, 2020
* Prefix logging messages with [TRT] (#497) * Add local build instructions (#498) * fix duplicate layer names bug (#446) (#467) Suppose we have a network with (not all distinct) layer names layer layer_1 layer When ImporterContext sees "layer", it sees it's not in mLayerNameCounts, and sets mLayerNameCounts["layer"] = 1 and adds a TRT layer with name "layer". It then sees "layer_1", concludes it's not in mLayerNameCounts, so it sets mLayerNameCounts["layer_1"] = 1 and adds a TRT layer with name "layer_1". NOW when it sees "layer", it sees that mLayerNameCounts["layer"] == 1, so we produce a "uniqueName" of "layer" + "_" + std::to_string(mLayerNameCounts["layer"] ), ie "layer_1", which is a name conflict for the TRT net. This change keeps track of all inserted names in a set and in the case of duplicates, tries suffix-appended modifications of the duplicated name by ever increasing integers until a name appears which has not been used. * Support Dynamic and 3D instanceNormalization (#515) * Add restrictions on multi-input convolutions (#521) * Update resize limitations in TensorRT (#538) Signed-off-by: Kevin Chen <kevinch@nvidia.com> Co-authored-by: Thomas Peters <thomas.d.peters@gmail.com>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
No description provided.