remove the codec name from the decoder definition in -d to only have the backend and the acceleration framework

The information in the command line is redundant as we pass the codec decoder along the codec test suite. We could rely only on codec test  suite to select the right decoder.


For example when you run:

`$ ./fluster.py run -s -d FFmpeg-H.265-Vulkan -ts JCT-VC-HEVC_V1`

We could directly run:

`$ ./fluster.py run -s -d FFmpeg-Vulkan -ts JCT-VC-HEVC_V1` 

as the test suite is h265 only.
