Qualcomm AI Engine Direct - Requantization Mechanism Implementation (#2823)
Summary: - Implement requantization so mix quantization ( e.g., 8bit quant + 16 bit quant) can be properly delegated to QNN. - Reusing test_qnn_backend_view_permute_matmul unit test to check mix quantization is working as expected. - Added etdump logic back to qnn_executor_runner that was deleted unintentionally during this PR: https://github.com/pytorch/executorch/commit/a531ca5a5ab5b5e503739a40a7cbec2e46767f0b#diff-f3647de74042ac9a417e2d4000a6f2db00c22c89fd028e9433d3c79ffb7d56f6 - Refactor common arguments in VIT. Pull Request resolved: https://github.com/pytorch/executorch/pull/2823 Reviewed By: kirklandsign Differential Revision: D55705289 Pulled By: cccclai fbshipit-source-id: c01744ce05f0ab27cf25fc28856ce150462470d5
W
winskuo-quic committed
61ad48d9c364cca312be0487214a96f4e15dbd44
Parent: 86b326a
Committed by Facebook GitHub Bot <facebook-github-bot@users.noreply.github.com>
on 4/7/2024, 10:27:52 PM