tvm.micro

MicroTVM module for bare-metal backends

Exceptions:

SessionTerminatedError	Raised when a transport read operation discovers that the remote session is terminated.
UnsupportedInModelLibraryFormatError	Raised when export_model_library_format does not support the given Module tree.

Classes:

AutoTvmModuleLoader(template_project_dir[, ...])	MicroTVM AutoTVM Module Loader
GeneratedProject(api_client, options)	Defines a glue interface to interact with a generated project through the API server.
Session([transport_context_manager, ...])	MicroTVM Device Session
TemplateProject(api_client)	Defines a glue interface to interact with a template project through the API Server.
TransportLogger(name, child[, logger, level])	Wraps a Transport implementation and logs traffic to the Python logging infrastructure.

Functions:

autotvm_build_func()	A dummy build function which causes autotvm to use a different export format.
copy_crt_config_header(platform, output_path)	Copy crt_config header file for a platform to destinatin.
create_local_aot_executor(session)	Create a local AoT executor driving execution on the remote CPU device given.
create_local_debug_executor(graph_json_str, ...)	Create a local debug runtime driving execution on the remote CPU device given.
create_local_graph_executor(graph_json_str, ...)	Create a local graph executor driving execution on the remote CPU device given.
export_model_library_format(mods, file_name)	Export the build artifact in Model Library Format.
generate_project(template_project_dir, ...)	Generate a project for an embedded platform that contains the given model.
get_microtvm_template_projects(platform)	Find microTVM template project directory for specific platform.
get_standalone_crt_dir()	Find the standalone_crt directory.

exception tvm.micro.SessionTerminatedError

Raised when a transport read operation discovers that the remote session is terminated.

exception tvm.micro.UnsupportedInModelLibraryFormatError

Raised when export_model_library_format does not support the given Module tree.

class tvm.micro.AutoTvmModuleLoader(template_project_dir, project_options=None, project_dir=None, use_existing=False)

MicroTVM AutoTVM Module Loader

Parameters

template_project_dirUnion[os.PathLike, str]

project template path

project_optionsdict

project generation option

project_dir: str

if use_existing is False: The path to save the generated microTVM Project. if use_existing is True: The path to a generated microTVM Project for debugging.

use_existing: bool

skips the project generation and opens transport to the project at the project_dir address.

参数:

• template_project_dir (PathLike | str)

• project_options (dict)

• project_dir (PathLike | str)

• use_existing (bool)

class tvm.micro.GeneratedProject(api_client, options)

Defines a glue interface to interact with a generated project through the API server.

class tvm.micro.Session(transport_context_manager=None, session_name='micro-rpc', timeout_override=None)

MicroTVM Device Session

Parameters

configdict

configuration for this session (as generated by tvm.micro.device.host.default_config(), for example)

Example

c_mod = ...  # some module generated with "c" as the target
dev_config = micro.device.arm.stm32f746xx.default_config('127.0.0.1', 6666)
with tvm.micro.Session(dev_config) as sess:
    micro_mod = sess.create_micro_mod(c_mod)

Methods:

__enter__()	Initialize this session and establish an RPC session with the on-device RPC server.
__exit__(exc_type, exc_value, exc_traceback)	Tear down this session and associated RPC session resources.
__init__([transport_context_manager, ...])	Configure a new session.

__enter__()

Initialize this session and establish an RPC session with the on-device RPC server.

Returns

Session :

Returns self.

__exit__(exc_type, exc_value, exc_traceback)

Tear down this session and associated RPC session resources.

__init__(transport_context_manager=None, session_name='micro-rpc', timeout_override=None)

Configure a new session.

Parameters

transport_context_managerContextManager[transport.Transport]

If given, flasher and binary should not be given. On entry, this context manager should establish a transport between this TVM instance and the device.

session_namestr

Name of the session, used for debugging.

timeout_overrideTransportTimeouts

If given, TransportTimeouts that govern the way Receive() behaves. If not given, this is determined by calling has_flow_control() on the transport.

class tvm.micro.TemplateProject(api_client)

Defines a glue interface to interact with a template project through the API Server.

Methods:

_check_project_options(options)	Check if options are valid ProjectOptions
generate_project(graph_executor_factory, ...)	Generate a project given GraphRuntimeFactory.
generate_project_from_mlf(...)	Generate a project from MLF file.

_check_project_options(options)

Check if options are valid ProjectOptions

参数:

options (dict)

generate_project(graph_executor_factory, project_dir, options)

Generate a project given GraphRuntimeFactory.

generate_project_from_mlf(model_library_format_path, project_dir, options)

Generate a project from MLF file.

参数:

options (dict)

class tvm.micro.TransportLogger(name, child, logger=None, level=20)

Wraps a Transport implementation and logs traffic to the Python logging infrastructure.

Methods:

close()	Release resources associated with this transport.
open()	Open any resources needed to send and receive RPC protocol data for a single session.
read(n, timeout_sec)	Read up to n bytes from the transport.
timeouts()	Return TransportTimeouts suitable for use with this transport.
write(data, timeout_sec)	Write data to the transport channel.

close()

Release resources associated with this transport.

open()

Open any resources needed to send and receive RPC protocol data for a single session.

read(n, timeout_sec)

Read up to n bytes from the transport.

Parameters

nint

Maximum number of bytes to read from the transport.

timeout_secUnion[float, None]

Number of seconds to wait for all n bytes to be received before timing out. The transport can wait additional time to account for transport latency or bandwidth limitations based on the selected configuration and number of bytes being received. If timeout_sec is 0, read should attempt to service the request in a non-blocking fashion. If timeout_sec is None, read should block until at least 1 byte of data can be returned.

Returns

bytes :

Data read from the channel. Less than n bytes may be returned, but 0 bytes should never be returned. If returning less than n bytes, the full timeout_sec, plus any internally-added timeout, should be waited. If a timeout or transport error occurs, an exception should be raised rather than simply returning empty bytes.

Raises

TransportClosedError :

When the transport layer determines that the transport can no longer send or receive data due to an underlying I/O problem (i.e. file descriptor closed, cable removed, etc).

IoTimeoutError :

When timeout_sec elapses without receiving any data.

timeouts()

Return TransportTimeouts suitable for use with this transport.

See the TransportTimeouts documentation in python/tvm/micro/session.py.

write(data, timeout_sec)

Write data to the transport channel.

Parameters

databytes

The data to write over the channel.

timeout_secUnion[float, None]

Number of seconds to wait for at least one byte to be written before timing out. The transport can wait additional time to account for transport latency or bandwidth limitations based on the selected configuration and number of bytes being received. If timeout_sec is 0, write should attempt to service the request in a non-blocking fashion. If timeout_sec is None, write should block until at least 1 byte of data can be returned.

Returns

int :

The number of bytes written to the underlying channel. This can be less than the length of data, but cannot be 0 (raise an exception instead).

Raises

TransportClosedError :

When the transport layer determines that the transport can no longer send or receive data due to an underlying I/O problem (i.e. file descriptor closed, cable removed, etc).

IoTimeoutError :

When timeout_sec elapses without receiving any data.

tvm.micro.autotvm_build_func()

A dummy build function which causes autotvm to use a different export format.

tvm.micro.copy_crt_config_header(platform, output_path)

Copy crt_config header file for a platform to destinatin.

Parameters

platformstr

Platform type which should be defined in MicroTVMTemplateProject.

output_path: Path

Output path for crt_config header file.

参数:

• platform (str)

• output_path (Path)

tvm.micro.create_local_aot_executor(session)

Create a local AoT executor driving execution on the remote CPU device given.

Parameters

sessionSession

A microTVM device session.

Returns

tvm.runtime.executor.aot_executor.AotModule :

A local AoT executor instance that executes on the remote device.

参数:

session (Session)

tvm.micro.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

graph_json_strstr

A string containing the graph representation.

modtvm.runtime.Module

The remote module containing functions in graph_json_str.

devicetvm.runtime.Device

The remote CPU execution device.

dump_rootOptional[str]

If given, passed as dump_root= to GraphModuleDebug.

Returns

tvm.contrib.GraphExecutor :

A local graph executor instance that executes on the remote device.

tvm.micro.create_local_graph_executor(graph_json_str, mod, device)

Create a local graph executor driving execution on the remote CPU device given.

Parameters

graph_json_strstr

A string containing the graph representation.

modtvm.runtime.Module

The remote module containing functions in graph_json_str.

devicetvm.runtime.Device

The remote CPU execution device.

Returns

tvm.contrib.GraphExecutor :

A local graph executor instance that executes on the remote device.

tvm.micro.export_model_library_format(mods, file_name)

Export the build artifact in Model Library Format.

This function creates a .tar archive containing the build artifacts in a standardized layout. It’s intended to allow downstream automation to build TVM artifacts against the C runtime.

Parameters

modExportableModule, List[ExportableModule]

The return value of tvm.build or tvm.relay.build.

file_namestr

Path to the .tar archive to generate.

Returns

file_namestr

The path to the generated .tar archive.

参数:

• mods (OperatorModule | AOTExecutorFactoryModule | GraphExecutorFactoryModule | List[OperatorModule | AOTExecutorFactoryModule | GraphExecutorFactoryModule])

• file_name (str | Path)

tvm.micro.generate_project(template_project_dir, module, generated_project_dir, options=None)

Generate a project for an embedded platform that contains the given model.

Parameters

template_project_pathpathlib.Path or str

Path to a template project containing a microTVM Project API server.

generated_project_pathpathlib.Path or str

Path to a directory to be created and filled with the built project.

moduleExportableModule

A runtime.Module exportable as Model Library Format. The value returned from tvm.relay.build or tvm.build.

optionsdict

If given, Project API options given to the microTVM API server found in both template_project_path and generated_project_path.

Returns

GeneratedProject :

A class that wraps the generated project and which can be used to further interact with it.

参数:

• template_project_dir (Path | str)

• module (OperatorModule | AOTExecutorFactoryModule | GraphExecutorFactoryModule)

• generated_project_dir (Path | str)

• options (dict)

tvm.micro.get_microtvm_template_projects(platform)

Find microTVM template project directory for specific platform.

Parameters

platformstr

Platform type which should be defined in MicroTVMTemplateProject.

Returns

str :

Path to template project directory for platform.

参数:

platform (str)

返回类型:

str

tvm.micro.get_standalone_crt_dir()

Find the standalone_crt directory.

Though the C runtime source lives in the tvm tree, it is intended to be distributed with any binary build of TVM. This source tree is intended to be integrated into user projects to run models targeted with –runtime=c.

Returns

str :

The path to the standalone_crt

返回类型:

str