[Mlir-commits] [mlir] [linalg] Add quantized version of `conv_3d_ncdhw_fcdhw` (PR #113953)

Javed Absar llvmlistbot at llvm.org
Wed Nov 13 04:35:15 PST 2024


================
@@ -4024,6 +4024,145 @@ structured_op: !LinalgStructuredOpConfig
                 - !ScalarExpression
                   scalar_arg: K
 --- !LinalgOpConfig
+metadata: !LinalgOpMetadata
+  name: conv_3d_ncdhw_fcdhw_q
+  cpp_class_name: Conv3DNcdhwFcdhwQOp
+  doc: |-
+    Performs 3-D convolution with zero point offsets.
+
+    Numeric casting is performed on the operands to the inner multiply, promoting
----------------
javedabsar1 wrote:

s/is performed on the operands to the inner multiply/
  is performed on the operands of the inner multiply ?

https://github.com/llvm/llvm-project/pull/113953


More information about the Mlir-commits mailing list