[MLAS] DequantizeLinear int8/uint8 #24818
sse2 dequantize linear
55384718
Merge branch 'main' into adrianl/dequantize-mlas
19157200
DQ NEON intrinsics
108ca0d9
Merge branch 'main' into adrianl/dequantize-mlas
8f190f0b
Rework SSE 2 to use a loop that loads 16 inputs at a time
b71efa07
adrianlizarraga
changed the title [MLAS] DequantizeLinear int8/uint8 [MLAS] DequantizeLinear int8/uint8 (multithreaded and vectorized) 249 days ago
adrianlizarraga
changed the title [MLAS] DequantizeLinear int8/uint8 (multithreaded and vectorized) [MLAS] DequantizeLinear int8/uint8 249 days ago
Disable new dequantize unit test for webgpu because webgpu requires t…
a288bcb1
Merge branch 'main' into adrianl/dequantize-mlas
73101067
Only enable multithreaded DQ for client/on-device workloads
f590ab0b
Mark thread_pool as unused
99e21927
typo fix
a547ee91
Update onnxruntime/core/providers/cpu/quantization/quantize_linear.cc
cda95458
jywu-msft
approved these changes
on 2025-07-09
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub