it downstream (when targeting a different environment, like TensorFlow lite for
example).
The MLIR module is expected to have a main() function that will be used as an
entry point. The inputs to the operations will be passed as argument to the
main() function and the returned values of the main function mapped to the
outputs.
Example usage:

```
import tensorflow as tf

        [`tf.compat.v1.distribute.experimental.ParameterServerStrategy`].
    *   Added `tf.distribute.experimental.coordinator.*` namespace, including
        the main API `ClusterCoordinator` for coordinating the training cluster,
        the related data structure `RemoteValue` and `PerWorkerValue`.

		{"WriteMetaDir", Func, 20},
	},
	"runtime/debug": {
		{"(*BuildInfo).String", Method, 18},
		{"BuildInfo", Type, 12},
		{"BuildInfo.Deps", Field, 12},
		{"BuildInfo.GoVersion", Field, 18},
		{"BuildInfo.Main", Field, 12},
		{"BuildInfo.Path", Field, 12},
		{"BuildInfo.Settings", Field, 18},
		{"BuildSetting", Type, 18},
		{"BuildSetting.Key", Field, 18},
		{"BuildSetting.Value", Field, 18},

            "type": "string"
          },

Class <org.gradle.language.nativeplatform.internal.MacroFunction> is not annotated (directly or via its package) with @org.gradle.api.NonNullApi in (MacroFunction.java:0)
Class <org.gradle.language.nativeplatform.internal.Names$Main> is not annotated (directly or via its package) with @org.gradle.api.NonNullApi in (Names.java:0)