Comment by dannyw

Building distributed training pipelines or optimising your ML stack (examples called out in the model card) isn’t harmful.