tree ae11f55f6bac939a51d5182eae441d322efb3e0e
parent 9272f8b9050096f39796227c5d89ed7b9905146d
author Narumol Prangnawarat <narumol.prangnawarat@arm.com> 1585214443 +0000
committer Narumol Prangnawarat <narumol.prangnawarat@arm.com> 1585239415 +0000

IVGCVSW-4597 Modify BF16 optimizer to Convert only inputs and weights of
Convolution2d and FullyConnected layers

 * Add InsertConvertFp32ToBf16LayersBefore
 * Add ConvertWeight to ConvertFp32NetworkToBf16Impl for Conv2d and FullyConnected
 * Allow different input and output when input is BF16 and output is FP32
Conv2d and FullyConnected layers
 * Unit tests

Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Ic8f92ff28edcae08a72a3114a28f50c4619f919b