I suggest we document something a bit more verbose: "Flag to enable or disable MKLDNN cache. It is on by default to optimize for performance. Should be disabled if your model has variable input shapes, as cache size may grow unbounded."
[ Full content available at: https://github.com/apache/incubator-mxnet/pull/12746 ] This message was relayed via gitbox.apache.org for devnull@infra.apache.org
