Commit f266b07
authored
[PT2E][X86] Add Inductor fusion passes of float8 qconv for X86Inductor backend (#3261)
* [Inductor][float8] Register qconv weight prepack pass for float8
* [Inductor][float8] Register qconv-unary fusion pass for float8
* [Inductor][float8] Register qconv-binary fusion pass for float8
* add comments1 parent d355d1f commit f266b07
File tree
2 files changed
+350
-128
lines changed- test/quantization/pt2e
- torchao/quantization/pt2e/inductor_passes
2 files changed
+350
-128
lines changed
0 commit comments