Fix paddle setup (#42254)

* expose api

* ref clipgradbynorm

* update

* Update __init__.py

python/paddle/incubate/distributed/models/moe/__init__.py

@@ -11,3 +11,8 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+
+from .gate import GShardGate, BaseGate, SwitchGate, NaiveGate
+from .moe_layer import MoELayer
+from .grad_clip import ClipGradForMOEByGlobalNorm
+ClipGradByGlobalNorm = ClipGradForMOEByGlobalNorm

python/setup.py.in

@@ -312,6 +312,8 @@ packages=['paddle',
           'paddle.distributed.auto_parallel.tuner',
           'paddle.distributed.auto_parallel.cost',
           'paddle.distributed.passes',
+          'paddle.distributed.models',
+          'paddle.distributed.models.moe',
           'paddle.framework',
           'paddle.jit',
           'paddle.jit.dy2static',
@@ -366,6 +368,10 @@ packages=['paddle',
           'paddle.incubate.nn.functional',
           'paddle.incubate.nn.layer',
           'paddle.incubate.optimizer.functional',
+          'paddle.incubate.distributed',
+          'paddle.incubate.distributed.models',
+          'paddle.incubate.distributed.models.moe',
+          'paddle.incubate.distributed.models.moe.gate',
           'paddle.io',
           'paddle.optimizer',
           'paddle.nn',