onnxruntime
[Quant tool] Handle input models with pre-quantized weights
#22633
Merged
Merged
[Quant tool] Handle input models with pre-quantized weights #22633
adrianlizarraga
merged
13 commits into
main
from
adrianl/quant-tool-handle-prequantized-weights
First working version that handles pre-quantized weights
e104ef44
Remove extra options; just handle pre-quantized weights if they exist
7af7f3a8
Undo import changes to run unittests locally
33b0266e
Use temp files to store unittest models
b653ccc2
Merge branch 'main' into adrianl/quant-tool-handle-prequantized-weights
5d65b4c6
Merge main and fix conflicts
29b839b9
Check for pre-quantized tensor instead of pre-quantized initializer s…
e87b53fb
Update unit tests for pre-quantized weights
eb8981eb
Merge branch 'main' into adrianl/quant-tool-handle-prequantized-weights
7adcbd70
Check that the scale/zero-point do not change for pre-quantized weigh…
0f0a3e0b
import annotations in onnx_model.py
b70729d7
Simplify and reduce diff
5096e7ee
Merge branch 'main' into adrianl/quant-tool-handle-prequantized-weights
3df2ddb5
baijumeswani
approved these changes
on 2024-11-14
Reviewers
Assignees
No one assigned
Labels
release:1.20.0
Milestone
No milestone
Login to write a write a comment.
Login via GitHub