fix(quantize): Overhaul INT8 static quantization to use QDQ format, unlocking TensorRT/NPU hardware acceleration#307
Open
jay7-tech wants to merge 2 commits intoopencv:mainfrom
Conversation
…rch64 hardware to resolve quantization paradox
…nlocking TensorRT/NPU hardware acceleration
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Here Problem
When generating INT8 ONNX models using the quantize-ort.py script, edge hardware accelerators (TensorRT, TIM-VX, CUDA NPUs) silently fallback to scalar CPU execution.
The script explicitly forces the legacy
QuantFormat.QOperator. This permanently blocks dynamic operator fusion on modern execution providers.Solution
QDQ(QuantizeLinear/DequantizeLinear) is currently the industry-standard layout required for these accelerators to compile fused INT8 engines. I refactored the pipeline to dynamically map and output QDQ tensor graphs (quant_format=QuantFormat.QDQ), which instantly unblocks native hardware acceleration for edge endpoints.