Fix Triton mini batch bug #688
Closed
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
… mapping back to the same source row in the dataframe. ex: model max batch = 4 seq_ids = [0, 0, 0, 1, 1, 2, 3, 3] Would cause the inference inputs for dataframe row 1 to occurr in two non-adjacent batches. This is a non-ideal fix as this doesn't handle the case where the inference input is so large that it spans more rows than the model's max batch size.
…sed on the shape to a method in TensorUtils
Merged
mdemoret-nv
requested changes
Feb 16, 2023
rapids-bot bot
pushed a commit
that referenced
this pull request
Feb 22, 2023
* Remove usage of `tensorShape_t` which was deprecated, and later removed. * Replace usage of tensor constructor in favor of the recommended `make_tensor` helper method. * Adds more C++ unittests * RMMTensor marked as a public symbol so the C++ tests can use it * Add `cuda-nvtx` package to CI driver build, needed for matx-0.3.0 Includes changes from PR #688 fixes #317 Authors: - David Gardner (https://github.com/dagardner-nv) Approvers: - Michael Demoret (https://github.com/mdemoret-nv) URL: #667
|
PR #667 contained this code and was merged. |
jjacobelli
pushed a commit
to jjacobelli/Morpheus
that referenced
this pull request
Mar 7, 2023
* Remove usage of `tensorShape_t` which was deprecated, and later removed. * Replace usage of tensor constructor in favor of the recommended `make_tensor` helper method. * Adds more C++ unittests * RMMTensor marked as a public symbol so the C++ tests can use it * Add `cuda-nvtx` package to CI driver build, needed for matx-0.3.0 Includes changes from PR nv-morpheus#688 fixes nv-morpheus#317 Authors: - David Gardner (https://github.com/dagardner-nv) Approvers: - Michael Demoret (https://github.com/mdemoret-nv) URL: nv-morpheus#667
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.
on_nextmethod of theInferenceClientStagewhich was getting a bit long and hard to follow, moving some logic to their own methods.Fixes #680