Search results

(topic/tracker) faceswap pipeline performance

Status: Open.

#112 In nod-ai/sdxl-scripts;

· monorimet opened on Jan 13, 2025

Attention to better MLPerf and beyond

Status: Open.

#108 In nod-ai/sdxl-scripts;

· raikonenfnu opened on Nov 19, 2024

Attention regression in ToM compared to MLPerf branch

Status: Open.

#107 In nod-ai/sdxl-scripts;

· MaheshRavishankar opened on Oct 15, 2024

[matmul] Look at LLVM instruction scheduling issues

Status: Open.

#92 In nod-ai/sdxl-scripts;

· antiagainst opened on Aug 5, 2024

·  

SDXL-MLPerf

[attention] Investigate overlapping matmul and softmax

Status: Open.

#91 In nod-ai/sdxl-scripts;

· antiagainst opened on Aug 5, 2024

·  

SDXL-MLPerf

[kernel] Fix excessive index calculation logic

Status: Open.

#90 In nod-ai/sdxl-scripts;

· antiagainst opened on Aug 5, 2024

·  

SDXL-MLPerf

(PUNET/I8): gfx1100 compilation issue with qbmm

Status: Open.

#44 In nod-ai/sdxl-scripts;

· monorimet opened on Jul 2, 2024

Failed executable compilation (tensor.extract_slice and an illegal vector.transfer_read)

sdxl-int8Issues replated to SDXL quantized model supportIssues replated to SDXL quantized model support

Status: Open.

#42 In nod-ai/sdxl-scripts;

· stellaraccident opened on Jun 30, 2024

[conv] Fix conv configuration like matmul to see through fused dequant ops

sdxl-int8Issues replated to SDXL quantized model supportIssues replated to SDXL quantized model support

Status: Open.

#39 In nod-ai/sdxl-scripts;

· antiagainst opened on Jun 29, 2024