Install Docker, then, from the VMAF directory, run:
docker build -t vmaf .
And to use it, just run:
docker run --rm vmaf [CLI]
Where [CLI]
is one of [run_vmaf, vmafossexec, vmaf_rc, run_vmaf_training, run_vmaf_in_batch, run_testing]
.
For example, if you are under root, to run run_vmaf
on a sample reference/distorted video pair under resource/yuv
:
docker run --rm -v $(pwd):/files vmaf \
run_vmaf yuv420p 576 324 \
/files/src01_hrc00_576x324.yuv \
/files/src01_hrc01_576x324.yuv \
--out-fmt json
Note that you need to first download the test videos from vmaf_resource.
wget https://github.com/Netflix/vmaf_resource/raw/master/python/test/resource/yuv/src01_hrc00_576x324.yuv
wget https://github.com/Netflix/vmaf_resource/raw/master/python/test/resource/yuv/src01_hrc01_576x324.yuv
To run vmafossexec
with a specified model file:
docker run --rm -v $(pwd):/files vmaf \
vmafossexec yuv420p 576 324 \
/files/src01_hrc00_576x324.yuv \
/files/src01_hrc01_576x324.yuv \
/files/model/vmaf_float_v0.6.1.pkl \
--log /dev/stdout