You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2020/12/09 18:17:19 UTC
[GitHub] [tvm] comaniac commented on a change in pull request #7070: Add autoscheduler support to tvmc
comaniac commented on a change in pull request #7070:
URL: https://github.com/apache/tvm/pull/7070#discussion_r539522022
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
Review comment:
It would be better to use an argument group for all hardware parameters.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
Review comment:
It might be confusing to have both `tune` and `autoschedule` subcommands. I would suggest either changing `tune` to `autotune` to be consistent with autoschedule`, or put auto-schedule as a mode in `tune` command.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
+
+ if args.rpc_tracker:
+
+ runner = auto_scheduler.RPCRunner(
+ key=args.rpc_key,
+ host=rpc_hostname,
+ port=rpc_port,
+ number=args.number,
+ repeat=args.repeat,
+ n_parallel=args.parallel,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+ else:
+ logger.info("starting localhost tuning")
+ runner = auto_scheduler.LocalRunner(
+ number=args.number,
+ repeat=args.repeat,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+
+ # Create the autoscheduler tuning options
+ tuning_options = auto_scheduler.TuningOptions(
+ num_measure_trials=args.trials,
+ measure_callbacks=[auto_scheduler.RecordToFile(args.output)],
+ runner=runner,
+ builder="local",
+ early_stopping=args.early_stopping,
+ )
+
+ # Specify hardware parameters
+ hardware_params = HardwareParams(
+ args.num_cores, args.vector_unit_bytes, args.cache_line_bytes, None, None, None, None, None
+ )
+
+ # Extract the tasks from the model
+ tasks, weights = get_tuning_tasks(
+ mod, params, target, target_host, args.desired_layout, hardware_params
+ )
+
+ # Schedule the tasks (i.e., produce a schedule for each task)
+ schedule_tasks(
+ tasks,
+ weights,
+ tuning_options,
+ args.tuning_records,
+ )
+
+
+def get_tuning_tasks(
+ mod, params, target, target_host=None, alter_layout=None, hardware_params=None
+):
+ """Get the tuning tasks for a given relay module.
+
+ Parameters
+ ----------
+ mod : tvm.relay.Module
+ The relay module from which to extract tuning tasks.
+ params : dict
+ The params for the relay module.
+ target : tvm.target.Target
+ The compilation target.
+ target_host : str, optional
+ The compilation target for the host.
+ alter_layout : str, optional
+ The layout to convert the graph to. Note, the convert layout
+ pass doesn't currently guarantee the whole of the graph will
+ be converted to the chosen layout.
+
+ Returns
+ -------
+ tasks : list of autotvm.Tasks
+ list of tasks to be tuned
+ weights : List[int]
+ the weight (i.e. the number of appearance) of extracted tasks
+ """
+ if alter_layout:
+ mod = common.convert_graph_layout(mod, alter_layout)
+
+ # Extract the tasks
+ tasks, task_weights = auto_scheduler.extract_tasks(
+ mod["main"], params, target=target, target_host=target_host, hardware_params=hardware_params
+ )
+
+ return tasks, task_weights
+
+
+def schedule_tasks(
+ tasks,
+ task_weights,
+ tuning_options,
+ tuning_records=None,
+):
+ """Generate the schedules for the different tasks (i.e., subgraphs) contained in the module.
+ Store the schedules in a json file that will be used later by the compiler.
+
+ Parameters
+ ----------
+ tasks : list
+ A list of autotvm.Tasks to tune.
+ task_weights : list
+ The weight (i.e. the number of appearance) of extracted tasks
+ tuning_records : str, optional
+ The json file used to preload the autoscheduler
+ tuning_options:
+ The options of tuning
Review comment:
Wrong order.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
+
+ if args.rpc_tracker:
+
+ runner = auto_scheduler.RPCRunner(
+ key=args.rpc_key,
+ host=rpc_hostname,
+ port=rpc_port,
+ number=args.number,
+ repeat=args.repeat,
+ n_parallel=args.parallel,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+ else:
+ logger.info("starting localhost tuning")
+ runner = auto_scheduler.LocalRunner(
+ number=args.number,
+ repeat=args.repeat,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+
+ # Create the autoscheduler tuning options
+ tuning_options = auto_scheduler.TuningOptions(
+ num_measure_trials=args.trials,
+ measure_callbacks=[auto_scheduler.RecordToFile(args.output)],
+ runner=runner,
+ builder="local",
+ early_stopping=args.early_stopping,
+ )
+
+ # Specify hardware parameters
+ hardware_params = HardwareParams(
+ args.num_cores, args.vector_unit_bytes, args.cache_line_bytes, None, None, None, None, None
+ )
+
+ # Extract the tasks from the model
+ tasks, weights = get_tuning_tasks(
+ mod, params, target, target_host, args.desired_layout, hardware_params
+ )
+
+ # Schedule the tasks (i.e., produce a schedule for each task)
+ schedule_tasks(
+ tasks,
+ weights,
+ tuning_options,
+ args.tuning_records,
+ )
+
+
+def get_tuning_tasks(
+ mod, params, target, target_host=None, alter_layout=None, hardware_params=None
+):
+ """Get the tuning tasks for a given relay module.
+
+ Parameters
+ ----------
+ mod : tvm.relay.Module
+ The relay module from which to extract tuning tasks.
+ params : dict
+ The params for the relay module.
+ target : tvm.target.Target
+ The compilation target.
+ target_host : str, optional
+ The compilation target for the host.
+ alter_layout : str, optional
+ The layout to convert the graph to. Note, the convert layout
+ pass doesn't currently guarantee the whole of the graph will
+ be converted to the chosen layout.
+
+ Returns
+ -------
+ tasks : list of autotvm.Tasks
+ list of tasks to be tuned
+ weights : List[int]
+ the weight (i.e. the number of appearance) of extracted tasks
+ """
+ if alter_layout:
+ mod = common.convert_graph_layout(mod, alter_layout)
+
+ # Extract the tasks
+ tasks, task_weights = auto_scheduler.extract_tasks(
+ mod["main"], params, target=target, target_host=target_host, hardware_params=hardware_params
+ )
+
+ return tasks, task_weights
+
+
+def schedule_tasks(
+ tasks,
+ task_weights,
+ tuning_options,
+ tuning_records=None,
+):
+ """Generate the schedules for the different tasks (i.e., subgraphs) contained in the module.
+ Store the schedules in a json file that will be used later by the compiler.
+
+ Parameters
+ ----------
+ tasks : list
+ A list of autotvm.Tasks to tune.
Review comment:
```suggestion
A list of auto_scheduler.SearchTask to tune.
```
##########
File path: python/tvm/driver/tvmc/common.py
##########
@@ -36,6 +36,93 @@ class TVMCException(Exception):
"""TVMC Exception"""
+def add_tuning_options(parser):
Review comment:
This function name might be confusing, as "tuning option" usually refers to the options for measurement. Maybe `add_tuning_cmds` or something like this would be clearer.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
Review comment:
This part is the same as `tune` and could be extracted as well.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
+
+ if args.rpc_tracker:
+
Review comment:
Remove this line.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
+
+ if args.rpc_tracker:
+
+ runner = auto_scheduler.RPCRunner(
+ key=args.rpc_key,
+ host=rpc_hostname,
+ port=rpc_port,
+ number=args.number,
+ repeat=args.repeat,
+ n_parallel=args.parallel,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+ else:
+ logger.info("starting localhost tuning")
+ runner = auto_scheduler.LocalRunner(
+ number=args.number,
+ repeat=args.repeat,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+
+ # Create the autoscheduler tuning options
+ tuning_options = auto_scheduler.TuningOptions(
+ num_measure_trials=args.trials,
+ measure_callbacks=[auto_scheduler.RecordToFile(args.output)],
+ runner=runner,
+ builder="local",
Review comment:
No need to provide the default value.
##########
File path: python/tvm/driver/tvmc/autoscheduler.py
##########
@@ -0,0 +1,212 @@
+# 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.
+"""
+Provides support to auto-tuning networks using AutoScheduler.
+"""
+import logging
+
+from urllib.parse import urlparse
+
+from tvm import auto_scheduler
+from tvm.auto_scheduler.auto_schedule import HardwareParams
+
+from . import common, frontends
+from .common import add_tuning_options
+from .main import register_parser
+
+
+# pylint: disable=invalid-name
+logger = logging.getLogger("TVMC")
+
+
+@register_parser
+def add_autoscheduler_parser(subparsers):
+ """ Include parser for 'autoschedule' subcommand """
+ parser = subparsers.add_parser("autoschedule", help="auto-schedule a model")
+ parser.set_defaults(func=drive_autoschedule)
+ add_tuning_options(parser)
+
+ parser.add_argument(
+ "--cache-line-bytes",
+ default=64,
+ help="the size of cache line in bytes",
+ )
+ parser.add_argument(
+ "--num-cores",
+ default=4,
+ help="the number of device cores",
+ )
+ parser.add_argument(
+ "--vector-unit-bytes",
+ default=16,
+ help="the width of vector units in bytes",
+ )
+ parser.add_argument(
+ "--model-format",
+ choices=frontends.get_frontend_names(),
+ help="specify input model format",
+ )
+
+
+def drive_autoschedule(args):
+ """Invoke auto-scheduling with command line arguments
+
+ Parameters
+ ----------
+ args: argparse.Namespace
+ Arguments from command line parser.
+ """
+
+ # extra arguments validation before importing the model, so that obvious errors
+ # are pointed in advance.
+ if args.rpc_tracker:
+ parsed_url = urlparse("//%s" % args.rpc_tracker)
+ rpc_hostname = parsed_url.hostname
+ rpc_port = parsed_url.port or 9090
+ logger.info("RPC tracker hostname: %s", rpc_hostname)
+ logger.info("RPC tracker port: %s", rpc_port)
+
+ if not args.rpc_key:
+ raise common.TVMCException(
+ "need to provide an RPC tracker key (--rpc-key) for remote tuning"
+ )
+
+ target = common.target_from_cli(args.target)
+ mod, params = frontends.load_model(args.FILE, args.model_format)
+
+ # min_repeat_ms should be:
+ # a. the value provided by the user, if any, or
+ # b. 0ms in case target is "cpu"; otherwise 1000ms
+ if args.min_repeat_ms is not None:
+ min_repeat_ms = args.min_repeat_ms
+ else:
+ min_repeat_ms = 0 if target.keys[0] == "cpu" else 1000
+ logger.debug("Default --min-repeat-ms for this target is %s", min_repeat_ms)
+
+ if args.rpc_tracker:
+
+ runner = auto_scheduler.RPCRunner(
+ key=args.rpc_key,
+ host=rpc_hostname,
+ port=rpc_port,
+ number=args.number,
+ repeat=args.repeat,
+ n_parallel=args.parallel,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+ else:
+ logger.info("starting localhost tuning")
+ runner = auto_scheduler.LocalRunner(
+ number=args.number,
+ repeat=args.repeat,
+ timeout=args.timeout,
+ min_repeat_ms=min_repeat_ms,
+ )
+
+ # Create the autoscheduler tuning options
+ tuning_options = auto_scheduler.TuningOptions(
+ num_measure_trials=args.trials,
+ measure_callbacks=[auto_scheduler.RecordToFile(args.output)],
+ runner=runner,
+ builder="local",
+ early_stopping=args.early_stopping,
+ )
+
+ # Specify hardware parameters
+ hardware_params = HardwareParams(
+ args.num_cores, args.vector_unit_bytes, args.cache_line_bytes, None, None, None, None, None
+ )
+
+ # Extract the tasks from the model
+ tasks, weights = get_tuning_tasks(
+ mod, params, target, target_host, args.desired_layout, hardware_params
+ )
+
+ # Schedule the tasks (i.e., produce a schedule for each task)
+ schedule_tasks(
+ tasks,
+ weights,
+ tuning_options,
+ args.tuning_records,
+ )
+
+
+def get_tuning_tasks(
Review comment:
This function looks unnecessary, and it's a bit weird to perform layout conversion in a function called "get_tuning_tasks".
##########
File path: tests/python/driver/tvmc/test_autoscheduler.py
##########
@@ -0,0 +1,101 @@
+# 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.
+import json
+import pytest
+import os
+import tarfile
+
+from os import path
+
+from tvm import auto_scheduler
+from tvm.driver import tvmc
+
+
+def _get_tasks(model):
+ mod, params = tvmc.frontends.load_model(model)
+ tasks, weights = tvmc.autoscheduler.get_tuning_tasks(mod, params, "llvm")
+ return (tasks, weights)
+
+
+def _autoscheduler_test_helper(
Review comment:
I'm not sure if we need to actually run auto_scheduler in this test. It seems to me that this helper is used by the following 4 test cases. However, TVMC should not be in charge of testing "tasks", "tuning records", "early stopping" and "no tuning records", especially those tests don't even use TVMC commands.
##########
File path: python/tvm/driver/tvmc/compiler.py
##########
@@ -82,6 +82,11 @@ def add_compile_parser(subparsers):
help="path to an auto-tuning log file by AutoTVM. If not presented, "
"the fallback/tophub configs will be used",
)
+ parser.add_argument(
+ "--use-autoscheduler",
+ action="store_true",
+ help="use the autoscheduler to generate the compute schedules",
+ )
Review comment:
You actually don't need users to tell you if auto_scheduler is used. Since AutoTVM and AutoScheduler have different log formats, you can simply use the following logic to know which one you should use:
```python
def is_auto_scheduler_log(log_file):
try:
auto_scheduler.load_records(log_file)
except Exception:
return False
return True
```
----------------------------------------------------------------
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