Add scale option to ToDtype. Remove ConvertDtype. #736

Job	Run time
unittests-extended / linux-job	17m 35s
onnx / linux-job	5m 6s
unittests-linux (3.10, linux.12xlarge, cpu) / linux-job	47s
unittests-linux (3.11, linux.12xlarge, cpu) / linux-job	38s
unittests-linux (3.8, linux.12xlarge, cpu) / linux-job	38s
unittests-linux (3.8, linux.g5.4xlarge.nvidia.gpu, cuda, 11.8) / linux-job	33m 20s
unittests-linux (3.9, linux.12xlarge, cpu) / linux-job	33s
unittests-macos (3.10, macos-12) / macos-job	1m 30s
unittests-macos (3.11, macos-12) / macos-job	1m 45s
unittests-macos (3.8, macos-12) / macos-job	1m 41s
unittests-windows (3.10, windows.4xlarge, cpu) / windows-job	1m 27s
unittests-macos (3.9, macos-12) / macos-job	1m 34s
unittests-macos (3.8, macos-m1-12) / macos-job	25m 31s
unittests-windows (3.11, windows.4xlarge, cpu) / windows-job	1m 34s
unittests-windows (3.8, windows.4xlarge, cpu) / windows-job	1m 16s
unittests-windows (3.8, windows.g5.4xlarge.nvidia.gpu, cuda, 11.8) / windows-job	26m 0s
unittests-windows (3.9, windows.4xlarge, cpu) / windows-job	1m 46s
	2h 2m 41s

Provide feedback