Fix bug of fp16 converter about the cast node and topology in sub-graph #291
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
|
Please add some descriptions so we know what kind of issue this PR aims to address. |
fatcat-z
reviewed
Jun 5, 2024
fatcat-z
reviewed
Jun 5, 2024
fatcat-z
reviewed
Jun 5, 2024
fatcat-z
reviewed
Jun 5, 2024
fatcat-z
reviewed
Jun 5, 2024
fatcat-z
reviewed
Jun 6, 2024
fatcat-z
reviewed
Jun 6, 2024
xiaowuhu
changed the title
xiaowuhu
changed the title
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Outdated suggestions cannot be applied.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.
There have a lot of bugs in the float16 converter after ORT 1.17 released, because the optimization rule changed in ORT. And customers also raised a lot of fp16 converter issue, which to be considered as corner case, but also need to fix. The original implementation is hard to understand, less maintainable. So I rewrite the function
convert_float_to_float16, the logic is simple to understand, and easy to modify each components.Also added 3 new features:
identitynode when it is unnecessary.castnode when it is unnecessary.the topology is not in order, the input of node 'x' is not output of any previous node.