projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: d3a6dbc) | patch
Add quant.const_fake_quant_per_axis op
authorFeng Liu <fengliuai@google.com>
Mon, 9 Sep 2019 22:42:07 +0000 (15:42 -0700)
committerA. Unique TensorFlower <gardener@tensorflow.org>
Mon, 9 Sep 2019 22:42:37 +0000 (15:42 -0700)
commitf4ae4762bf7d64d7ca46d05206955c9b44cedc49
treee1a5dc72eedc496474a62a2def3d03f089acfaf6tree | snapshot
parentd3a6dbc0b895bdfdae3627c00d35066c4f51b032commit | diff
Add quant.const_fake_quant_per_axis op

Comparing to the existing quant.const_fake_quant op, the min and max attributes
of this new op is for each channel of last dimension of the input.

PiperOrigin-RevId: 268093722
mlir/include/mlir/Dialect/QuantOps/QuantOps.td diff | blob | history
mlir/test/Dialect/QuantOps/parse-ops.mlir diff | blob | history
Domain: System / Toolchain;