Update qkeras converter for array and auto po2 scale #135
Conversation
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
added 13 commits
March 22, 2024 15:23
…rn new rounding modes.
Fixed rounding_mode specifier in convert_quantized_bits
Commented out assertion on non-representability
…ay_and_auto_po2_scale Merging changes from rounding_mode_new.
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Outdated suggestions cannot be applied.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.
This PR updates the QKeras converter to enable it to convert neural networks with tensor scale factors, and with scale factors set to "auto_po2".
Previously the converter added a "Quant" node for every "quantized_bits" function in qkeras. Now it adds an additional "Mul" node after a Quant node. This is because of differences in how QKeras and QONNX handle dequantization.
This additional "Mul" node might be a problem for downstream onnx uses, but can be removed when scale factors are equal to one using a transform.
This PR "follows" after PR: Rounding mode new #134 which introduces additional rounding modes. This is why it contains rounding mode changes for now (since they are not yet merged into main).