-
Notifications
You must be signed in to change notification settings - Fork 136
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.
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
Decouple TritonInferenceStage from pipeline mode #1402
Merged
rapids-bot
merged 58 commits into
nv-morpheus:branch-24.03
from
dagardner-nv:triton-logits-1378
Jan 8, 2024
Merged
Decouple TritonInferenceStage from pipeline mode #1402
rapids-bot
merged 58 commits into
nv-morpheus:branch-24.03
from
dagardner-nv:triton-logits-1378
Jan 8, 2024
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
dagardner-nv
added
breaking
Breaking change
improvement
Improvement to existing functionality
labels
Dec 1, 2023
drobison00
approved these changes
Dec 1, 2023
mdemoret-nv
approved these changes
Dec 11, 2023
…onstructor we need to specify one anytime we are using the other pipeline mode
dagardner-nv
changed the title
Make logits a constructor argument to TritonInferenceStage
Decouple TritonInferenceStage from pipeline mode
Dec 11, 2023
mdemoret-nv
requested changes
Dec 13, 2023
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Make the following changes:
- Remove
AE
support from the triton stage all together - Remove all implementations of the worker class
- Make all options of the worker class instance values (need_logits, and inout_mapping)
- Set defaults for inout_mapping similar to default for
need_logits
(i.e. by mode if not supplied)
…riton-logits-1378 [no ci]
…ge from the AE pipeline [no ci]
…the parent constructor, bypassing the checks in the MultiTensorMessage class. Manually construct a valid seq_ids array still need to detmine why the inference stage is emitting an inference message and not a response message [no ci]
…riton-logits-1378
mdemoret-nv
approved these changes
Jan 8, 2024
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Lets get these changes merged but make an issue before you do to try and remove the inference worker class entirely.
2 tasks
/merge |
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.
Description
needs_logits
&inout_mapping
a constructor arguments toTritonInferenceStage
.TritonInferenceWorker
classneeds_logits
&inout_mapping
arguments default to default values inferred by the pipeline mode.examples/log_parsing
example as a subclass ofTritonInferenceStage
removing several lines of redundant code.examples/log_parsing
example now works in C++ modeinf-triton
from the AE mode in the CLICloses #1378
By Submitting this PR I confirm: