def add_logger_settings(script, args):
# Always required since it is used to print the exit message.
script.append_preimport("from polygraphy.logger import G_LOGGER")
logger_settings = []
verbosity_count = args_util.get(args, "verbose")
if verbosity_count >= 4:
logger_settings.append("G_LOGGER.severity = G_LOGGER.ULTRA_VERBOSE")
elif verbosity_count == 3:
logger_settings.append("G_LOGGER.severity = G_LOGGER.SUPER_VERBOSE")
elif verbosity_count == 2:
logger_settings.append("G_LOGGER.severity = G_LOGGER.EXTRA_VERBOSE")
elif verbosity_count == 1:
logger_settings.append("G_LOGGER.severity = G_LOGGER.VERBOSE")
if args_util.get(args, "silent"):
logger_settings.append("G_LOGGER.severity = G_LOGGER.CRITICAL")
log_format = misc.default_value(args_util.get(args, "log_format"), [])
for fmt in args.log_format:
if fmt == "no-colors":
logger_settings.append("G_LOGGER.colors = False")
elif fmt == "timestamp":
logger_settings.append("G_LOGGER.timestamp = True")
elif fmt == "line-info":
logger_settings.append("G_LOGGER.line_info = True")
for setting in logger_settings:
script.append_preimport(setting)
def add_onnx_loader(script, args, disable_outputs=None, suffix=None):
if args_util.get(args, "model_type") == "onnx":
script.add_import(imports=["OnnxFromPath"], frm="polygraphy.backend.onnx")
loader_str = Script.invoke("OnnxFromPath", args_util.get(args, "model_file"))
loader_name = script.add_loader(loader_str, "load_onnx", suffix=suffix)
else:
G_LOGGER.verbose("Attempting to load as a TensorFlow model, using TF2ONNX to convert to ONNX. "
"If this is not correct, please specify --model-type", mode=LogMode.ONCE)
script.add_import(imports=["OnnxFromTfGraph"], frm="polygraphy.backend.onnx")
loader_str = Script.invoke("OnnxFromTfGraph", add_tf_loader(script, args, disable_outputs=True, suffix=suffix),
opset=args_util.get(args, "opset"), fold_constant=False if args_util.get(args, "no_const_folding") else None)
loader_name = script.add_loader(loader_str, "export_onnx_from_tf", suffix=suffix)
modify_onnx_str = get_modify_onnx_str(script, args, loader_name, disable_outputs=disable_outputs)
if modify_onnx_str is not None:
loader_name = script.add_loader(modify_onnx_str, "modify_onnx")
save_onnx = args_util.get(args, "save_onnx")
SAVE_ONNX = "SaveOnnx"
save_onnx_str = Script.invoke(SAVE_ONNX, loader_name, path=save_onnx)
if save_onnx_str != Script.invoke(SAVE_ONNX, loader_name):
script.add_import(imports=[SAVE_ONNX], frm="polygraphy.backend.onnx")
loader_name = script.add_loader(save_onnx_str, "save_onnx")
return loader_name
def add_trt_config_loader(script, args, data_loader_name):
profiles = []
for (min_shape, opt_shape, max_shape) in args_util.get(args, "profiles"):
profile_str = "Profile()"
for name in min_shape.keys():
profile_str += Script.format_str(".add({:}, min={:}, opt={:}, max={:})", name, min_shape[name], opt_shape[name], max_shape[name])
profiles.append(Inline(profile_str))
if profiles:
script.add_import(imports=["Profile"], frm="polygraphy.backend.trt")
sep = Inline("\n{:}".format(constants.TAB))
profiles = Script.format_str("[{:}{:}\n]", sep, Inline((",{:}".format(sep)).join(profiles)))
profile_name = script.add_loader(profiles, "profiles")
else:
profile_name = None
calibrator = None
if args_util.get(args, "int8"):
script.add_import(imports=["DataLoader"], frm="polygraphy.comparator")
script.add_import(imports=["Calibrator"], frm="polygraphy.backend.trt")
calibrator = Script.invoke("Calibrator", data_loader=Inline(data_loader_name) if data_loader_name else Inline("DataLoader()"),
cache=args_util.get(args, "calibration_cache"))
config_loader_str = Script.invoke_if_nondefault("CreateTrtConfig", max_workspace_size=args_util.get(args, "workspace"), tf32=args_util.get(args, "tf32"),
fp16=args_util.get(args, "fp16"), int8=args_util.get(args, "int8"), strict_types=args_util.get(args, "strict_types"),
profiles=profile_name, calibrator=Inline(calibrator) if calibrator else None)
if config_loader_str is not None:
script.add_import(imports=["CreateConfig as CreateTrtConfig"], frm="polygraphy.backend.trt")
config_loader_name = script.add_loader(config_loader_str, "create_trt_config")
else:
config_loader_name = None
return config_loader_name
def add_serialized_onnx_loader(script, args, disable_outputs=None):
model_file = args_util.get(args, "model_file")
needs_modify = get_modify_onnx_str(script, args, "check_needs_modify", disable_outputs) is not None
should_import_raw = args_util.get(args, "model_type") == "onnx" and not needs_modify
if should_import_raw:
script.add_import(imports=["BytesFromPath"], frm="polygraphy.backend.common")
onnx_loader = script.add_loader(Script.invoke("BytesFromPath", model_file), "load_serialized_onnx")
else:
script.add_import(imports=["BytesFromOnnx"], frm="polygraphy.backend.onnx")
onnx_loader = add_onnx_loader(script, args, disable_outputs=disable_outputs)
onnx_loader = script.add_loader(Script.invoke("BytesFromOnnx", onnx_loader), "serialize_onnx")
return onnx_loader
def add_trt_runner(script, args, data_loader_name):
script.add_import(imports=["TrtRunner"], frm="polygraphy.backend.trt")
if args.model_type == "engine":
loader_name = tool_util.add_trt_serialized_engine_loader(script, args)
else:
script.add_import(imports=["EngineFromNetwork"],
frm="polygraphy.backend.trt")
loader_name = tool_util.add_trt_network_loader(script, args)
config_loader_name = tool_util.add_trt_config_loader(
script, args, data_loader_name=data_loader_name)
loader_str = Script.invoke("EngineFromNetwork",
loader_name,
config=config_loader_name)
loader_name = script.add_loader(loader_str, "build_engine")
SAVE_ENGINE = "SaveEngine"
save_engine = Script.invoke(SAVE_ENGINE,
loader_name,
path=args_util.get(args, "save_engine"))
if save_engine != Script.invoke(SAVE_ENGINE, loader_name):
script.add_import(imports=[SAVE_ENGINE], frm="polygraphy.backend.trt")
loader_name = script.add_loader(save_engine, "save_engine")
script.add_runner(Script.invoke("TrtRunner", loader_name))
def _wrap_if_plugins(script, args, obj_name):
plugins = args_util.get(args, "plugins")
if plugins:
script.add_import(imports=["LoadPlugins"], frm="polygraphy.backend.trt")
loader_str = Script.invoke("LoadPlugins", obj_name, plugins=plugins)
obj_name = script.add_loader(loader_str, "load_plugins")
return obj_name
def add_tf_config_loader(script, args):
config_loader_str = Script.invoke_if_nondefault("CreateConfig", gpu_memory_fraction=args_util.get(args, "gpu_memory_fraction"),
allow_growth=args_util.get(args, "allow_growth"), use_xla=args_util.get(args, "xla"))
if config_loader_str is not None:
script.add_import(imports=["CreateConfig"], frm="polygraphy.backend.tf")
config_loader_name = script.add_loader(config_loader_str, "create_tf_config")
else:
config_loader_name = None
return config_loader_name
def add_trt_legacy_runner(script, args):
script.add_import(imports=["TrtLegacyRunner"],
frm="polygraphy.backend.trt_legacy")
G_LOGGER.warning(
"Legacy TensorRT runner only supports implicit batch TensorFlow/UFF, ONNX, and Caffe models"
)
if args.model_type == "onnx":
script.add_import(imports=["ParseNetworkFromOnnxLegacy"],
frm="polygraphy.backend.trt_legacy")
onnx_loader = tool_util.add_onnx_loader(script,
args,
disable_outputs=True)
loader_name = script.add_loader(
Script.format_str("ParseNetworkFromOnnxLegacy({:})", onnx_loader),
"parse_network_from_onnx_legacy")
elif args.model_type == "caffe":
script.add_import(imports=["LoadNetworkFromCaffe"],
frm="polygraphy.backend.trt_legacy")
loader_name = script.add_loader(
Script.format_str("LoadNetworkFromCaffe({:}, {:}, {:}, {:})",
args.model_file, args.caffe_model,
args.trt_outputs, args.batch_size),
"parse_network_from_caffe")
else:
script.add_import(imports=["LoadNetworkFromUff"],
frm="polygraphy.backend.trt_legacy")
if args.model_type == "uff":
script.add_import(imports=["LoadUffFile"],
frm="polygraphy.backend.trt_legacy")
shapes = {name: shape for name, (_, shape) in args.inputs.items()}
loader_name = script.add_loader(
Script.format_str("LoadUffFile({:}, {:}, {:})",
args.model_file,
misc.default_value(shapes, {}),
args.trt_outputs), "load_uff_file")
else:
script.add_import(imports=["ConvertToUff"],
frm="polygraphy.backend.trt_legacy")
loader_name = script.add_loader(
Script.format_str(
"ConvertToUff({:}, save_uff={:}, preprocessor={:})",
tool_util.add_tf_loader(script, args), args.save_uff,
args.preprocessor), "convert_to_uff")
loader_name = script.add_loader(
Script.format_str("LoadNetworkFromUff({:}, uff_order={:})",
loader_name, args.uff_order),
"uff_network_loader")
runner_str = Script.format_str(
"TrtLegacyRunner({:}, {:}, {:}, fp16={:}, tf32={:}, load_engine={:}, save_engine={:}, layerwise={:}, plugins={:})",
loader_name, args.workspace, args.batch_size, args.fp16, args.tf32,
args.model_file if args.model_type == "engine" else None,
args.save_engine,
args_util.get(args, "trt_outputs") == constants.MARK_ALL, args.plugins)
script.add_runner(runner_str)
def get_modify_onnx_str(script, args, loader_name, disable_outputs=None):
if disable_outputs:
outputs = None
exclude_outputs = None
else:
outputs = _get_outputs_arg(script, args, "onnx_outputs")
exclude_outputs = args_util.get(args, "onnx_exclude_outputs")
if hasattr(args, "shape_inference"):
do_shape_inference = args_util.get(args, "shape_inference")
else:
do_shape_inference = None if args_util.get(args, "no_shape_inference") else True
MODIFY_ONNX = "ModifyOnnx"
modify_onnx_str = Script.invoke(MODIFY_ONNX, loader_name, do_shape_inference=do_shape_inference,
outputs=outputs, exclude_outputs=exclude_outputs)
if modify_onnx_str != Script.invoke(MODIFY_ONNX, loader_name):
script.add_import(imports=[MODIFY_ONNX], frm="polygraphy.backend.onnx")
return modify_onnx_str
return None
def add_trt_network_loader(script, args):
model_file = args_util.get(args, "model_file")
outputs = _get_outputs_arg(script, args, "trt_outputs")
if args_util.get(args, "network_api"):
CREATE_NETWORK_FUNC = Inline("create_network")
script.add_import(imports=["CreateNetwork"], frm="polygraphy.backend.trt")
script.add_import(imports=["extend"], frm="polygraphy.common.func")
script.append_prefix("# Manual TensorRT network creation")
script.append_prefix("@extend(CreateNetwork())")
script.append_prefix("def {:}(builder, network):".format(CREATE_NETWORK_FUNC))
script.append_prefix("{tab}import tensorrt as trt\n".format(tab=constants.TAB))
script.append_prefix("{tab}# Define your network here. Make sure to mark outputs!".format(tab=constants.TAB))
net_inputs = args_util.get(args, "inputs")
if net_inputs:
for name, (dtype, shape) in net_inputs.items():
script.append_prefix("{tab}{name} = network.add_input(name='{name}', shape={shape}, dtype=trt.float32) # TODO: Set dtype".format(
name=name, shape=shape, tab=constants.TAB))
script.append_prefix("{tab}# TODO: network.mark_output(...)\n".format(tab=constants.TAB))
return CREATE_NETWORK_FUNC
if args_util.get(args, "ext"):
script.add_import(imports=["NetworkFromOnnxPath"], frm="polygraphy.backend.trt")
loader_str = Script.invoke("NetworkFromOnnxPath", _wrap_if_plugins(script, args, model_file), explicit_precision=args_util.get(args, "explicit_precision"))
loader_name = script.add_loader(loader_str, "parse_network_from_onnx")
else:
script.add_import(imports=["NetworkFromOnnxBytes"], frm="polygraphy.backend.trt")
onnx_loader = add_serialized_onnx_loader(script, args, disable_outputs=True)
loader_str = Script.invoke("NetworkFromOnnxBytes", _wrap_if_plugins(script, args, onnx_loader), explicit_precision=args_util.get(args, "explicit_precision"))
loader_name = script.add_loader(loader_str, "parse_network_from_onnx")
MODIFY_NETWORK = "ModifyNetwork"
modify_network_str = Script.invoke(MODIFY_NETWORK, loader_name, outputs=outputs, exclude_outputs=args_util.get(args, "trt_exclude_outputs"))
if modify_network_str != Script.invoke(MODIFY_NETWORK, loader_name):
script.add_import(imports=[MODIFY_NETWORK], frm="polygraphy.backend.trt")
loader_name = script.add_loader(modify_network_str, "modify_network")
return loader_name
def find(self):
def run(indices):
self.mark_layers(indices)
return self.check_network("-".join(map(str, indices)))
# Finds num worst indices in acc_results
def find_worst(num, acc_results):
acc_mapping = list(acc_results.values())[0][
0] # First iteration of first runner-pair.
# Compute for each layer: atol / prev_atol, to determine which layers contribute the greatest error.
# It is not enough to simply find the max(atol), because that doesn't account for error introduced
# by previous layers.
items = list(acc_mapping.items())
ratios = []
for (_, prev_tols), (outname,
cur_tols) in zip(items[:-1], items[1:]):
ratio = cur_tols.required_atol / prev_tols.required_atol
ratios.append((ratio, outname))
# Mark more layers on each iteration
ratios = sorted(ratios, reverse=True)[:num]
G_LOGGER.verbose(
"Found worst {:} layers (Format: (error ratio, tensor name)): {:}"
.format(num, ratios))
return [output_mapping[outname] for (ratio, outname) in ratios]
if not args_util.get(self.args, "trt_outputs"):
G_LOGGER.critical(
"worst-first requires all outputs to be marked as network outputs mode to determine where errors are being introduced. "
"Please enable --trt-outputs mark all, and ensure that your golden outputs also include layer-wise results"
)
output_mapping = {
} # Maps output tensor names to producer layer indices
for layer_index, layer in enumerate(self.network):
for out_index in range(layer.num_outputs):
output_mapping[layer.get_output(out_index).name] = layer_index
indices = []
acc_results = run(indices)
max_outputs = len(list(acc_results.values())[0][0]) - 1
iter_num = 0
# indices will be at most one less than the number of layers, since we're comparing layers against subsequent ones.
while not bool(acc_results) and len(indices) < max_outputs:
iter_num += 1
indices = find_worst(self.args.top * iter_num, acc_results)
acc_results = run(indices)
if bool(acc_results):
return indices
def add_data_loader(script, args):
def omit_none_tuple(tup):
if all([elem is None for elem in tup]):
return None
return tup
int_range = omit_none_tuple(tup=(args_util.get(args, "int_min"), args_util.get(args, "int_max")))
float_range = omit_none_tuple(tup=(args_util.get(args, "float_min"), args_util.get(args, "float_max")))
input_metadata_str = Inline(repr(args_util.get(args, "inputs"))) if args_util.get(args, "inputs") else None
if input_metadata_str:
script.add_import(imports=["TensorMetadata"], frm="polygraphy.common")
data_loader = Script.invoke_if_nondefault("DataLoader", seed=args_util.get(args, "seed"), iterations=args_util.get(args, "iterations"),
input_metadata=input_metadata_str, int_range=int_range, float_range=float_range)
if data_loader is not None:
data_loader_name = Inline("data_loader")
script.add_import(imports=["DataLoader"], frm="polygraphy.comparator")
script.append_prefix(Script.format_str("\n# Inference Inputs Loader\n{:} = {:}\n", data_loader_name, Inline(data_loader)))
else:
data_loader_name = None
return data_loader_name
def add_comparator(script, args, data_loader_name, cmd_run):
script.add_import(imports=["Comparator"], frm="polygraphy.comparator")
script.add_import(imports=["sys"])
comparator_run = Script.invoke("Comparator.run",
script.get_runners(),
warm_up=args.warm_up,
data_loader=data_loader_name,
use_subprocess=args.use_subprocess)
script.append_suffix(
Script.format_str("\n# Runner Execution\nresults = {:}",
Inline(comparator_run)))
if args.load_results:
G_LOGGER.verbose("Will load runner results from: {:}".format(
args.load_results))
script.add_import(imports=["misc"], frm="polygraphy.util")
script.append_suffix(
Script.format_str(
"\n# Load results\nfor load_output in {:}:\n{:}results.update(misc.pickle_load(load_output))",
args.load_results, Inline(constants.TAB)))
if args.save_results:
G_LOGGER.verbose("Will save runner results to: {:}".format(
args.save_results))
script.add_import(imports=["misc"], frm="polygraphy.util")
script.append_suffix(
Script.format_str(
"\n# Save results\nmisc.pickle_save({:}, results)",
args.save_results))
top_k = args_util.get(args, "top_k")
if top_k is not None:
script.add_import(imports=["PostprocessFunc"],
frm="polygraphy.comparator")
script.append_suffix(
Script.format_str(
"\n# Postprocessing - Apply Top-{:}\nresults = Comparator.postprocess(results, PostprocessFunc.topk_func(k={:}))",
top_k, top_k))
script.append_suffix("\nsuccess = True")
if len(
args.runners
) > 1 or args.load_results: # Only do comparisons if there's actually something to compare.
script.append_suffix("# Accuracy Comparison")
compare_func_str = Script.invoke_if_nondefault(
"CompareFunc.basic_compare_func",
rtol=args.rtol,
atol=args.atol,
check_shapes=False if args.no_shape_check else None,
fail_fast=args.fail_fast)
compare_func = None
if compare_func_str:
script.add_import(imports=["CompareFunc"],
frm="polygraphy.comparator")
compare_func = "compare_func"
script.append_suffix(
Script.format_str("{:} = {:}", Inline(compare_func),
Inline(compare_func_str)))
compare_accuracy = Script.invoke("Comparator.compare_accuracy",
Inline("results"),
compare_func=Inline(compare_func)
if compare_func is not None else None,
fail_fast=args.fail_fast)
script.append_suffix(
Script.format_str("success &= bool({:})\n",
Inline(compare_accuracy)))
if args.validate:
script.append_suffix(
"# Validation\nsuccess &= Comparator.validate(results)\n")
if cmd_run is None:
cmd_run = Inline("' '.join(sys.argv)")
script.append_suffix(
Script.format_str(
'# Report Results\ncmd_run={cmd}\nif success:\n G_LOGGER.success("PASSED | Command: {{}}".format(cmd_run))\nelse:\n G_LOGGER.error("FAILED | Command: {{}}".format(cmd_run))',
cmd=cmd_run))
script.append_suffix("sys.exit(0 if success else 1)")
def add_tf_loader(script, args, disable_outputs=None, suffix=None):
if disable_outputs:
outputs = None
else:
outputs = _get_outputs_arg(script, args, "tf_outputs")
model_file = args_util.get(args, "model_file")
model_type = args_util.get(args, "model_type")
save_pb = args_util.get(args, "save_pb")
save_tensorboard = args_util.get(args, "save_tensorboard")
if model_type == "ckpt":
G_LOGGER.verbose("Loading a TensorFlow checkpoint. Please ensure you are not using the --use-subprocess flag".format(model_file), mode=LogMode.ONCE)
script.add_import(imports=["GraphFromCkpt"], frm="polygraphy.backend.tf")
loader_id = "load_ckpt"
loader_str = Script.invoke("GraphFromCkpt", model_file, args_util.get(args, "ckpt"))
elif model_type == "keras":
script.add_import(imports=["GraphFromKeras"], frm="polygraphy.backend.tf")
loader_id = "load_keras"
loader_str = Script.invoke("GraphFromKeras", model_file)
else:
script.add_import(imports=["GraphFromFrozen"], frm="polygraphy.backend.tf")
G_LOGGER.verbose("Attempting to load as a frozen graph. If this is not correct, please specify --model-type", mode=LogMode.ONCE)
loader_id = "load_frozen"
loader_str = Script.invoke("GraphFromFrozen", model_file)
loader_name = script.add_loader(loader_str, loader_id, suffix=suffix)
if args_util.get(args, "freeze_graph"):
script.add_import(imports=["OptimizeGraph"], frm="polygraphy.backend.tf")
loader_name = script.add_loader(Script.invoke("OptimizeGraph", loader_name), "optimize_graph", suffix=suffix)
if args_util.get(args, "tftrt"):
script.add_import(imports=["UseTfTrt"], frm="polygraphy.backend.tf")
loader_str = Script.invoke("UseTfTrt", loader_name, max_workspace_size=args_util.get(args, "workspace"), fp16=args_util.get(args, "fp16"), int8=args_util.get(args, "int8"),
max_batch_size=args_util.get(args, "batch_size"), is_dynamic_op=args_util.get(args, "dynamic_op"), minimum_segment_size=args_util.get(args, "minimum_segment_size"))
loader_name = script.add_loader(loader_str, "use_tftrt", suffix=suffix)
MODIFY_TF = "ModifyGraph"
modify_tf_str = Script.invoke(MODIFY_TF, loader_name, outputs=outputs)
if modify_tf_str != Script.invoke(MODIFY_TF, loader_name):
script.add_import(imports=[MODIFY_TF], frm="polygraphy.backend.tf")
loader_name = script.add_loader(modify_tf_str, "modify_tf")
engine_dir = None
if args_util.get(args, "tftrt"):
engine_dir = args_util.get(args, "save_engine")
WRITE_TF = "SaveGraph"
write_tf_str = Script.invoke(WRITE_TF, loader_name, path=save_pb, tensorboard_dir=save_tensorboard, engine_dir=engine_dir)
if write_tf_str != Script.invoke(WRITE_TF, loader_name):
script.add_import(imports=[WRITE_TF], frm="polygraphy.backend.tf")
loader_name = script.add_loader(write_tf_str, "save_tf")
return loader_name
def _get_outputs_arg(script, args, name):
outputs = args_util.get(args, name)
if outputs == constants.MARK_ALL:
outputs = Inline("constants.MARK_ALL")
script.add_import(["constants"], frm="polygraphy.common")
return outputs
def add_trt_serialized_engine_loader(script, args):
script.add_import(imports=["EngineFromBytes"], frm="polygraphy.backend.trt")
script.add_import(imports=["BytesFromPath"], frm="polygraphy.backend.common")
load_engine = script.add_loader(Script.invoke("BytesFromPath", args_util.get(args, "model_file")), "load_engine")
return script.add_loader(Script.invoke("EngineFromBytes", _wrap_if_plugins(script, args, load_engine)), "deserialize_engine")
