debug_runtime -> debug_executor

docs/dev/debugger.rst

@@ -134,12 +134,12 @@ How to use Debugger?
 
 3. In frontend script file instead of
    ``from tvm.contrib import graph_executor`` import the
-   ``debug_runtime``
-   ``from tvm.contrib.debugger import debug_runtime as graph_executor``
+   ``debug_executor``
+   ``from tvm.contrib.debugger import debug_executor as graph_executor``
 
 ::
 
-    from tvm.contrib.debugger import debug_runtime as graph_executor
+    from tvm.contrib.debugger import debug_executor as graph_executor
     m = graph_executor.create(graph, lib, dev, dump_root="/tmp/tvmdbg")
     # set inputs
     m.set_input('data', tvm.nd.array(data.astype(dtype)))

python/tvm/contrib/cuda_graph/cuda_graph_executor.py

@@ -131,4 +131,4 @@ def debug_get_output(self, node, out):
         out : NDArray
             The output array container
         """
-        raise NotImplementedError("Please use debugger.debug_runtime as graph_executor instead.")
+        raise NotImplementedError("Please use debugger.debug_executor as graph_executor instead.")

python/tvm/contrib/graph_executor.py

@@ -270,7 +270,7 @@ def debug_get_output(self, node, out):
         out : NDArray
             The output array container
         """
-        raise NotImplementedError("Please use debugger.debug_runtime as graph_executor instead.")
+        raise NotImplementedError("Please use debugger.debug_executor as graph_executor instead.")
 
     def load_params(self, params_bytes):
         """Load parameters from serialized byte array of parameter dict.

python/tvm/driver/tvmc/runner.py

@@ -27,7 +27,7 @@
 from tvm import rpc
 from tvm.autotvm.measure import request_remote
 from tvm.contrib import graph_executor as runtime
-from tvm.contrib.debugger import debug_runtime
+from tvm.contrib.debugger import debug_executor
 from tvm.relay import load_param_dict
 
 from . import common
@@ -370,7 +370,7 @@ def run_module(
 
         if profile:
             logger.debug("creating runtime with profiling enabled")
-            module = debug_runtime.create(graph, lib, dev, dump_root="./prof")
+            module = debug_executor.create(graph, lib, dev, dump_root="./prof")
         else:
             logger.debug("creating runtime with profiling disabled")
             module = runtime.create(graph, lib, dev)

python/tvm/micro/__init__.py

@@ -26,7 +26,7 @@
 from .model_library_format import export_model_library_format, UnsupportedInModelLibraryFormatError
 from .session import (
     create_local_graph_executor,
-    create_local_debug_runtime,
+    create_local_debug_executor,
     Session,
     SessionTerminatedError,
 )

python/tvm/micro/session.py

@@ -23,7 +23,7 @@
 from ..error import register_error
 from .._ffi import get_global_func
 from ..contrib import graph_executor
-from ..contrib.debugger import debug_runtime
+from ..contrib.debugger import debug_executor
 from ..rpc import RPCSession
 from .transport import IoTimeoutError
 from .transport import TransportLogger
@@ -218,7 +218,7 @@ def create_local_graph_executor(graph_json_str, mod, device):
     )
 
 
-def create_local_debug_runtime(graph_json_str, mod, device, dump_root=None):
+def create_local_debug_executor(graph_json_str, mod, device, dump_root=None):
     """Create a local debug runtime driving execution on the remote CPU device given.
 
     Parameters
@@ -242,7 +242,7 @@ def create_local_debug_runtime(graph_json_str, mod, device, dump_root=None):
     """
     device_type_id = [device.device_type, device.device_id]
     fcreate = get_global_func("tvm.graph_executor_debug.create")
-    return debug_runtime.GraphModuleDebug(
+    return debug_executor.GraphModuleDebug(
         fcreate(graph_json_str, mod, lookup_remote_linked_param, *device_type_id),
         [device],
         graph_json_str,

tests/python/unittest/test_runtime_graph_debug.py

@@ -28,7 +28,7 @@
 import numpy as np
 from tvm import rpc
 from tvm.contrib import utils
-from tvm.contrib.debugger import debug_runtime
+from tvm.contrib.debugger import debug_executor
 
 
 @tvm.testing.requires_llvm
@@ -75,7 +75,7 @@ def myadd(*args):
         mlib_proxy = tvm.support.FrontendTestModule()
         mlib_proxy["myadd"] = myadd
         try:
-            mod = debug_runtime.create(graph, mlib_proxy, tvm.cpu(0))
+            mod = debug_executor.create(graph, mlib_proxy, tvm.cpu(0))
         except ValueError:
             return
 
@@ -171,7 +171,7 @@ def check_remote():
         remote.upload(path_dso)
         mlib = remote.load_module("dev_lib.so")
         try:
-            mod = debug_runtime.create(graph, mlib, remote.cpu(0))
+            mod = debug_executor.create(graph, mlib, remote.cpu(0))
         except ValueError:
             print("Skip because debug runtime not enabled")
             return

tests/python/unittest/test_runtime_module_based_interface.py

@@ -19,7 +19,7 @@
 from tvm.relay import testing
 import tvm
 from tvm.contrib import graph_executor
-from tvm.contrib.debugger import debug_runtime
+from tvm.contrib.debugger import debug_executor
 from tvm.contrib.cuda_graph import cuda_graph_executor
 import tvm.testing
 
@@ -527,7 +527,7 @@ def test_debug_graph_executor():
     tvm.testing.assert_allclose(out, verify(data), atol=1e-5)
 
     # debug graph executor wrapper
-    debug_g_mod = debug_runtime.GraphModuleDebug(
+    debug_g_mod = debug_executor.GraphModuleDebug(
         complied_graph_lib["debug_create"]("default", dev),
         [dev],
         complied_graph_lib.get_json(),

vta/scripts/tune_resnet.py

@@ -29,7 +29,7 @@
 from tvm.autotvm.measure.measure_methods import request_remote
 from tvm.autotvm.tuner import XGBTuner, GATuner, RandomTuner, GridSearchTuner
 from tvm.contrib import graph_executor, utils, download
-from tvm.contrib.debugger import debug_runtime
+from tvm.contrib.debugger import debug_executor
 import vta
 from vta.testing import simulator
 from vta.top import graph_pack
@@ -325,7 +325,7 @@ def tune_tasks(
 
         # If detailed runtime info is needed build with debug runtime
         if opt.debug_profile:
-            m = debug_runtime.create(graph, lib, ctx)
+            m = debug_executor.create(graph, lib, ctx)
         else:
             m = graph_executor.create(graph, lib, ctx)
 

vta/tutorials/frontend/deploy_classification.py

@@ -53,7 +53,7 @@
 from tvm import te
 from tvm import rpc, autotvm, relay
 from tvm.contrib import graph_executor, utils, download
-from tvm.contrib.debugger import debug_runtime
+from tvm.contrib.debugger import debug_executor
 from tvm.relay import transform
 
 import vta