[GitHub] [incubator-tvm] electriclilies commented on a change in pull request #5962: [Ansor][AutoTVM v2.0] Part 0: Ansor minimum system for auto schedule generating

[GitBox <gi...@apache.org>]

electriclilies commented on a change in pull request #5962:
URL: https://github.com/apache/incubator-tvm/pull/5962#discussion_r450434208



##########
File path: python/tvm/ansor/compute_dag.py
##########
@@ -0,0 +1,141 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, 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.
+
+""" Computational graph and its analysis tools """
+
+import hashlib
+
+import tvm._ffi
+from tvm.runtime import Object
+from tvm.te import PlaceholderOp, ComputeOp
+
+from .loop_state import State, StateObject
+from .utils import get_const_tuple
+from .workload_registry import workload_key_to_tensors
+
+from . import _ffi_api
+
+
+@tvm._ffi.register_object("ansor.ComputeDAG")
+class ComputeDAG(Object):
+    """
+    Computation declaration graph.
+
+    Parameters
+    ----------
+    compute : Union[List[Tensor], str]
+        `Tensor`s or workload key for a compute declaration.
+    """
+    def __init__(self, compute):
+        if isinstance(compute, str):
+            compute = workload_key_to_tensors(compute)
+        elif isinstance(compute, list):
+            for item in compute:
+                if not isinstance(item, tvm.te.Tensor):
+                    raise ValueError("The input of ComputeDAG should be a list of Tensor")
+        else:
+            raise ValueError("Invalid compute: " + compute + ". Expect a string or list of Tensor")
+        self.__init_handle_by_constructor__(_ffi_api.ComputeDAG, compute)
+
+    def get_init_state(self):
+        """ Get init state of this ComputeDAG.
+
+        Returns
+        -------
+        state : State
+            The initial State without any transform steps.
+        """
+        return State(_ffi_api.ComputeDAGGetInitState(self), self)
+
+    def apply_steps_from_state(self, state):
+        """
+        Apply transform steps according to the history of a State.
+
+        Parameters
+        ----------
+        state : Union[State, StateObject]
+            The target state to be applied to TVM schedule.
+
+        Returns
+        -------
+            A `te.schedule` and the target `te.Tensor`s to be used in `tvm.lower` or `tvm.build`
+        """
+        state_obj = state if isinstance(state, StateObject) else state.state_object
+        return _ffi_api.ComputeDAGApplyStepsFromState(self, state_obj)
+
+    def print_python_code_from_state(self, state):

Review comment:
       I think the name is confusing too, I would suggest print_transform_history_from_state, or print_python_history_from_state, since the function prints the transform history in python.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org