未验证 提交 c23f09fe 编写于 作者: C Chen Weihang 提交者: GitHub

Support load state_dict from save_params/persistables (#27298)

* support load state_dict from save_params/persistables

* remove failed unittest

* add load eof check & unittest

* remove eof check
上级 c67c3916
......@@ -195,58 +195,11 @@ def load_dygraph(model_path, config=None):
params_file_path = model_prefix + ".pdparams"
opti_file_path = model_prefix + ".pdopt"
# deal with argument `configs`
configs = config
if configs is None:
configs = SaveLoadConfig()
if not os.path.exists(params_file_path) and not os.path.exists(
opti_file_path):
# Load state dict by `jit.save/io.save_inference_model` save format
# NOTE(chenweihang): [ Compatibility of save_inference_model save format ]
# The model saved by `save_inference_model` does not completely correspond to
# the information required by the `state_dict` under the dygraph.
# `save_inference_model` not save structured name, we need to remind
# the user to configure the `use_structured_name` argument when `set_state_dict`
# NOTE(chenweihang): `jit.save` doesn't save optimizer state
# 1. check model path
if not os.path.isdir(model_prefix):
raise ValueError("Model saved directory '%s' is not exists." %
model_prefix)
# deal with argument `config`
if config is None:
config = SaveLoadConfig()
# 2. load program desc & construct _ProgramHolder
programs = _construct_program_holders(model_path,
configs.model_filename)
# 3. load layer parameters & buffers
# NOTE: using fluid.dygraph.guard() here will cause import error in py2
with guard():
persistable_var_dict = _construct_params_and_buffers(
model_prefix,
programs,
configs.separate_params,
configs.params_filename,
append_suffix=False)
# 4. construct state_dict
para_dict = dict()
for var_name in persistable_var_dict:
para_dict[var_name] = persistable_var_dict[var_name].numpy()
# if __variables.info__ exists, we can recover structured_name
var_info_path = os.path.join(model_prefix, EXTRA_VAR_INFO_FILENAME)
if os.path.exists(var_info_path):
with open(var_info_path, 'rb') as f:
extra_var_info = pickle.load(f)
structured_para_dict = dict()
for var_name in para_dict:
structured_name = extra_var_info[var_name].get(
'structured_name', None)
assert structured_name is not None, "Cannot find saved variable (%s)'s structured name in saved model." % var_name
structured_para_dict[structured_name] = para_dict[var_name]
para_dict = structured_para_dict
else:
if os.path.exists(params_file_path) or os.path.exists(opti_file_path):
# Load state dict by `save_dygraph` save format
para_dict = {}
if os.path.exists(params_file_path):
......@@ -254,12 +207,103 @@ def load_dygraph(model_path, config=None):
para_dict = pickle.load(f) if six.PY2 else pickle.load(
f, encoding='latin1')
if not configs.keep_name_table and "StructuredToParameterName@@" in para_dict:
if not config.keep_name_table and "StructuredToParameterName@@" in para_dict:
del para_dict["StructuredToParameterName@@"]
if os.path.exists(opti_file_path):
with open(opti_file_path, 'rb') as f:
opti_dict = pickle.load(f) if six.PY2 else pickle.load(
f, encoding='latin1')
else:
# check model path
if not os.path.isdir(model_prefix):
raise ValueError("Model saved directory '%s' is not exists." %
model_prefix)
# check whether model file exists
if config.model_filename is None:
model_filename = '__model__'
else:
model_filename = config.model_filename
model_file_path = os.path.join(model_path, model_filename)
if os.path.exists(model_file_path):
# Load state dict by `jit.save/io.save_inference_model` save format
# NOTE(chenweihang): [ Compatibility of save_inference_model save format ]
# The model saved by `save_inference_model` does not completely correspond to
# the information required by the `state_dict` under the dygraph.
# `save_inference_model` not save structured name, we need to remind
# the user to configure the `use_structured_name` argument when `set_state_dict`
# NOTE(chenweihang): `jit.save` doesn't save optimizer state
# 1. load program desc & construct _ProgramHolder
programs = _construct_program_holders(model_path,
config.model_filename)
# 2. load layer parameters & buffers
# NOTE: using fluid.dygraph.guard() here will cause import error in py2
with guard():
persistable_var_dict = _construct_params_and_buffers(
model_prefix,
programs,
config.separate_params,
config.params_filename,
append_suffix=False)
# 3. construct state_dict
para_dict = dict()
for var_name in persistable_var_dict:
para_dict[var_name] = persistable_var_dict[var_name].numpy()
# if __variables.info__ exists, we can recover structured_name
var_info_path = os.path.join(model_prefix,
EXTRA_VAR_INFO_FILENAME)
if os.path.exists(var_info_path):
with open(var_info_path, 'rb') as f:
extra_var_info = pickle.load(f)
structured_para_dict = dict()
for var_name in para_dict:
structured_name = extra_var_info[var_name].get(
'structured_name', None)
assert structured_name is not None, "Cannot find saved variable (%s)'s structured name in saved model." % var_name
structured_para_dict[structured_name] = para_dict[
var_name]
para_dict = structured_para_dict
else:
# load state dict by `io.save_params/persistables` save format
# TODO(chenweihang): [ Now only supports loading parameters seperately ]
# If users save all parameters as one file, the [ variable.name -> variable ]
# mapping info will lost, so users need to give variable list, but users build
# variable list in dygraph mode is difficult, we recommend users to use
# paddle.io.load_program_state in this case
# Try to load all the files in the directory in VarBase format,
# the file name is used as the name of VarBase
load_var_list = []
# 1. load file names
var_name_list = []
for root, _, files in os.walk(model_path):
for filename in files:
file_path = os.path.join(root, filename)
tmp_var_name = os.path.relpath(file_path, model_path)
var_name = tmp_var_name.replace("\\", "/")
var_name_list.append(var_name)
# 2. create and load VarBase
with guard():
for name in var_name_list:
new_var = _varbase_creator(name=name, persistable=True)
_dygraph_tracer().trace_op(
type='load',
inputs={},
outputs={'Out': new_var},
attrs={'file_path': os.path.join(model_path, name)})
load_var_list.append(new_var)
# 3. construct state_dict
para_dict = dict()
for var in load_var_list:
para_dict[var.name] = var.numpy()
return para_dict, opti_dict
......@@ -64,7 +64,7 @@ class TestLoadStateDictFromSaveInferenceModel(unittest.TestCase):
self.batch_size = 128
self.batch_num = 10
def train_and_save_model(self):
def train_and_save_model(self, only_params=False):
with new_program_scope():
startup_program = fluid.default_startup_program()
main_program = fluid.default_main_program()
......@@ -102,11 +102,15 @@ class TestLoadStateDictFromSaveInferenceModel(unittest.TestCase):
static_param_dict[param.name] = fluid.executor._fetch_var(
param.name)
fluid.io.save_inference_model(
self.save_dirname, ["img"], [prediction],
exe,
model_filename=self.model_filename,
params_filename=self.params_filename)
if only_params:
fluid.io.save_params(
exe, self.save_dirname, filename=self.params_filename)
else:
fluid.io.save_inference_model(
self.save_dirname, ["img"], [prediction],
exe,
model_filename=self.model_filename,
params_filename=self.params_filename)
return static_param_dict
......@@ -120,9 +124,7 @@ class TestLoadStateDictFromSaveInferenceModel(unittest.TestCase):
self.params_filename = None
orig_param_dict = self.train_and_save_model()
configs = paddle.SaveLoadConfig()
configs.separate_params = True
load_param_dict, _ = paddle.load(self.save_dirname, configs)
load_param_dict, _ = paddle.load(self.save_dirname)
self.check_load_state_dict(orig_param_dict, load_param_dict)
def test_load_with_model_filename(self):
......@@ -160,6 +162,14 @@ class TestLoadStateDictFromSaveInferenceModel(unittest.TestCase):
load_param_dict, _ = paddle.load(self.save_dirname, configs)
self.check_load_state_dict(orig_param_dict, load_param_dict)
def test_load_state_dict_from_save_params(self):
self.save_dirname = "static_mnist.load_state_dict.save_params"
self.params_filename = None
orig_param_dict = self.train_and_save_model(True)
load_param_dict, _ = paddle.load(self.save_dirname)
self.check_load_state_dict(orig_param_dict, load_param_dict)
if __name__ == '__main__':
unittest.main()
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册