the consistency between trt and modelopt is quite different? #4364
Labels
triaged
Issue has been triaged by maintainers
waiting for feedback
Requires more information from user to make progress on the issue.
The int8 model output gap between trt and modelopt-quant-onnx inference is large.
Is this a known problem?
Is there a solution?
The text was updated successfully, but these errors were encountered: