Clara train covid19 ct lung segmentation error

Hi there, i am trying to utilise the clara_train_covid19_ct_lung_seg to initially perform inference on several COVID-19 positive CT scans. I have downloaded the Clara Train SDK cotainer and pulled the container from inside the Clara SDK container with the built-in NGC commands. At present i only want to use the pre-trained models for inferance, so am i correct in saying i use the infer.sh script for this as instructed by the Clara train docs. When doing this however i receive this error:

tensorflow.python.framework.errors_impl.InvalidArgumentError: NodeDef missing attr ‘reduce_op’ from Op<name=HorovodAllreduce; signature=tensor:T → sum:T; attr=T:type,allowed=[DT_INT32, DT_INT64, DT_HALF, DT_FLOAT, DT_DOUBLE]; attr=reduce_op:int>; NodeDef: {{node DistributedAdamOptimizer_Allreduce/HorovodAllreduce_mul_3_0}}

My (limited) experience would tell me this is a TensorFlow incompatibility issue. Have i done something wrong or am i overlooking something?

Your support is very much appreciated!