PR #22111 Address ZeroK case for Gemm for CPU and CUDA

Address ZeroK case for Gemm for CPU and CUDA #22111

yuslepukhin merged 12 commits into main from yuslepukhin/zeroK_Gemm

Address ZeroK case for Gemm for CPU and CUDA

c50c04d0

yuslepukhin requested a review from

snnn 1 year ago

yuslepukhin requested a review from

pranavsharma 1 year ago

yuslepukhin requested a review from

Lafi7e 1 year ago

yuslepukhin requested a review from

yufenglee 1 year ago

yuslepukhin requested a review from

tianleiwu 1 year ago

pranavsharma dismissed these changes on 2024-09-17

snnn commented on 2024-09-17

Add QGemm K == 0 handling

4713219f

yuslepukhin dismissed their stale review via 4713219f 1 year ago

Make GCC happy with .template<>()

aba1d52e

Address data conversion issues

5c336c77

Implemenet clipping properly

837fcf73

snnn dismissed these changes on 2024-09-19

yuslepukhin requested a review from

edgchen1 1 year ago

Rework BiasBroadcast

0353d5f6

yuslepukhin dismissed their stale review via 0353d5f6 1 year ago

Fix zp fill out

c68f2da5

Rework zp fill out

96017936

Make sure output is float with y_zp not present

9ca55b41

Account for b_scale being a vector of N

09f42ecc

Remove QGemm changes

af0d25e2

snnn dismissed these changes on 2024-09-19

yuslepukhin removed review request from

Lafi7e 1 year ago

yuslepukhin removed review request from

edgchen1 1 year ago

edgchen1 commented on 2024-09-19

Address review comments

2676a6c7

yuslepukhin dismissed their stale review via 2676a6c7 1 year ago

edgchen1 approved these changes on 2024-09-20

yuslepukhin merged fe8a10ca into main 1 year ago

yuslepukhin deleted the yuslepukhin/zeroK_Gemm branch 1 year ago

Reviewers

edgchen1

snnn

pranavsharma

yufenglee

tianleiwu

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

onnxruntime Address ZeroK case for Gemm for CPU and CUDA #22111 Merged

Address ZeroK case for Gemm for CPU and CUDA #22111

onnxruntime
Address ZeroK case for Gemm for CPU and CUDA
#22111

Merged