wehub-resource-sync
968 lines
34 KiB
Python
968 lines
34 KiB
Python
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# Licensed 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 sys
|
|
import unittest
|
|
|
|
import numpy as np
|
|
from op_test import get_device, get_device_place, is_custom_device
|
|
|
|
import paddle
|
|
from paddle import base
|
|
from paddle.base.framework import EagerParamBase
|
|
|
|
sys.path.append("../dygraph_to_static")
|
|
from dygraph_to_static_utils import enable_to_static_guard
|
|
|
|
|
|
class L1(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
self._param_attr = base.ParamAttr(
|
|
initializer=paddle.nn.initializer.Constant(value=0.1)
|
|
)
|
|
self.w1 = self.create_parameter(
|
|
attr=self._param_attr, shape=[2, 2], dtype='float32', is_bias=False
|
|
)
|
|
self.w2 = self.create_parameter(
|
|
attr=self._param_attr, shape=[2, 2], dtype='float32', is_bias=False
|
|
)
|
|
|
|
def forward(self):
|
|
return self.w1 + self.w2
|
|
|
|
|
|
class L2(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
self.layer1 = L1()
|
|
self.layer2 = L1()
|
|
|
|
def forward(self):
|
|
return self.layer1() + self.layer2()
|
|
|
|
|
|
class L3(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
self.layer1 = L2()
|
|
self.layer2 = L2()
|
|
|
|
def forward(self):
|
|
return self.layer1() + self.layer2()
|
|
|
|
|
|
class TestBaseLayer(unittest.TestCase):
|
|
def test_one_level(self):
|
|
l = L1()
|
|
ret = l()
|
|
expected_names = ['l1.w1', 'l1.w2']
|
|
idx = 0
|
|
for name, _ in l.named_parameters(prefix='l1'):
|
|
self.assertEqual(name, expected_names[idx])
|
|
idx += 1
|
|
np.testing.assert_allclose(
|
|
ret.numpy(), 0.2 * np.ones([2, 2]), rtol=1e-05
|
|
)
|
|
|
|
def test_three_level(self):
|
|
l = L3()
|
|
expected_names = [
|
|
'l3.layer1.layer1.w1',
|
|
'l3.layer1.layer1.w2',
|
|
'l3.layer1.layer2.w1',
|
|
'l3.layer1.layer2.w2',
|
|
'l3.layer2.layer1.w1',
|
|
'l3.layer2.layer1.w2',
|
|
'l3.layer2.layer2.w1',
|
|
'l3.layer2.layer2.w2',
|
|
]
|
|
idx = 0
|
|
for name, _ in l.named_parameters(prefix='l3'):
|
|
self.assertEqual(name, expected_names[idx])
|
|
idx += 1
|
|
ret = l()
|
|
np.testing.assert_allclose(
|
|
ret.numpy(), 0.8 * np.ones([2, 2]), rtol=1e-05
|
|
)
|
|
|
|
def test_add_parameter_with_error(self):
|
|
net = paddle.nn.Layer()
|
|
param = net.create_parameter(shape=[1])
|
|
|
|
with self.assertRaises(TypeError):
|
|
net.add_parameter(10, param)
|
|
|
|
with self.assertRaises(KeyError):
|
|
net.add_parameter("param.name", param)
|
|
|
|
with self.assertRaises(KeyError):
|
|
net.add_parameter("", param)
|
|
|
|
with self.assertRaises(KeyError):
|
|
net.test_param = 10
|
|
net.add_parameter("test_param", param)
|
|
|
|
with self.assertRaises(TypeError):
|
|
net.add_parameter("no_param", 10)
|
|
|
|
load_param = net.create_parameter(shape=[1])
|
|
net._loaddict_holder[load_param.name] = load_param
|
|
net.add_parameter("load_param", load_param)
|
|
|
|
|
|
class BufferLayer(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
buffer_var = paddle.to_tensor(np.zeros([2, 4]).astype('int32'))
|
|
self.register_buffer("layer_buffer", buffer_var)
|
|
|
|
def forward(self):
|
|
pass
|
|
|
|
|
|
class BufferNet(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
self.buffer_layer = BufferLayer()
|
|
self.w1 = self.create_parameter(
|
|
shape=[2, 2], dtype='float32', is_bias=False
|
|
)
|
|
buffer_var = paddle.to_tensor(np.ones([2, 4]).astype('int32'))
|
|
self.register_buffer("net_buffer", buffer_var)
|
|
|
|
self.new_buffer = paddle.to_tensor(np.ones([4, 2]).astype('int32'))
|
|
|
|
def forward(self):
|
|
pass
|
|
|
|
|
|
class TestBuffer(unittest.TestCase):
|
|
def test_buffers_and_named_buffers(self):
|
|
def names(named_buffers):
|
|
return [name for name, _ in named_buffers]
|
|
|
|
layer = BufferLayer()
|
|
net = BufferNet()
|
|
|
|
self.assertEqual(len(layer.buffers()), 1)
|
|
self.assertEqual(names(layer.named_buffers()), ['layer_buffer'])
|
|
|
|
self.assertEqual(len(net.buffers()), 3)
|
|
self.assertEqual(
|
|
names(net.named_buffers()),
|
|
['net_buffer', 'new_buffer', 'buffer_layer.layer_buffer'],
|
|
)
|
|
|
|
self.assertEqual(len(net.buffers(include_sublayers=False)), 2)
|
|
self.assertEqual(
|
|
names(net.named_buffers(include_sublayers=False)),
|
|
['net_buffer', 'new_buffer'],
|
|
)
|
|
|
|
def test_register_buffer_with_error(self):
|
|
net = paddle.nn.Layer()
|
|
var = paddle.to_tensor(np.zeros([1]))
|
|
|
|
with self.assertRaisesRegex(
|
|
TypeError, "name of buffer should be a string"
|
|
):
|
|
net.register_buffer(12, var)
|
|
|
|
with self.assertRaisesRegex(
|
|
TypeError, "buffer should be a Paddle.Tensor"
|
|
):
|
|
net.register_buffer(
|
|
"buffer_name", EagerParamBase([2, 2], 'float32')
|
|
)
|
|
|
|
with self.assertRaisesRegex(KeyError, "name of buffer can not contain"):
|
|
net.register_buffer("buffer.name", var)
|
|
|
|
with self.assertRaisesRegex(
|
|
KeyError, "name of buffer can not be empty"
|
|
):
|
|
net.register_buffer("", var)
|
|
|
|
net.attr_name = 10
|
|
with self.assertRaisesRegex(KeyError, "already exists"):
|
|
net.register_buffer("attr_name", var)
|
|
|
|
del net.attr_name
|
|
net.attr_name = EagerParamBase([2, 2], 'float32')
|
|
with self.assertRaisesRegex(KeyError, "already exists"):
|
|
net.register_buffer("attr_name", var)
|
|
|
|
def test_register_buffer_same_name(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
var2 = paddle.to_tensor(np.zeros([2]))
|
|
var3 = paddle.to_tensor(np.zeros([3]))
|
|
|
|
net.register_buffer("buffer_name", var1)
|
|
self.assert_var_base_equal(net.buffer_name, var1)
|
|
net.register_buffer("buffer_name", var2)
|
|
self.assert_var_base_equal(net.buffer_name, var2)
|
|
net.register_buffer("buffer_name", var3)
|
|
self.assert_var_base_equal(net.buffer_name, var3)
|
|
|
|
def test_buffer_not_persistable(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
self.assertEqual(len(net.buffers()), 1)
|
|
self.assertEqual(len(net.state_dict()), 0)
|
|
|
|
def test_buffer_not_persistable_del(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
del net.buffer_name
|
|
self.assertEqual(len(net.buffers()), 0)
|
|
|
|
def test_buffer_not_persistable_overwrite(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
var2 = paddle.to_tensor(np.zeros([2]))
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
net.register_buffer("buffer_name", var2)
|
|
|
|
# Allow to overwrite a non-persistable buffer with a persistable var.
|
|
self.assertEqual(len(net.buffers()), 1)
|
|
self.assertEqual(len(net.state_dict()), 1)
|
|
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
self.assertEqual(len(net.buffers()), 1)
|
|
self.assertEqual(len(net.state_dict()), 0)
|
|
|
|
def test_buffer_not_persistable_assign(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
|
|
# Assigning Nones will remove the buffer, but allow to re-assign
|
|
# to remark it as buffer.
|
|
net.buffer_name = None
|
|
self.assertEqual(len(net.buffers()), 0)
|
|
self.assertEqual(len(net.state_dict()), 0)
|
|
|
|
net.buffer_name = var1
|
|
self.assertEqual(len(net.buffers()), 1)
|
|
self.assertEqual(len(net.state_dict()), 0)
|
|
|
|
# Re-assign a EagerParamBase will remove the buffer.
|
|
net.buffer_name = EagerParamBase([2, 2], 'float32')
|
|
self.assertEqual(len(net.buffers()), 0)
|
|
self.assertEqual(len(net.state_dict()), 1)
|
|
|
|
def test_buffer_not_persistable_load(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([1]))
|
|
net.register_buffer("buffer_name", var1, persistable=False)
|
|
net.load_dict({})
|
|
|
|
def test_buffer_state_dict(self):
|
|
net = paddle.nn.Layer()
|
|
var1 = paddle.to_tensor(np.zeros([2, 3]))
|
|
var2 = paddle.to_tensor(np.zeros([3, 2]))
|
|
net.register_buffer("buffer_var1", var1)
|
|
net.register_buffer("buffer_var2", var2, persistable=False)
|
|
|
|
self.assertEqual(len(net.state_dict()), 1)
|
|
self.assertEqual(
|
|
[name for name, _ in net.state_dict().items()], ["buffer_var1"]
|
|
)
|
|
|
|
# load state_dict
|
|
net_load = paddle.nn.Layer()
|
|
var = paddle.to_tensor(np.ones([2, 3]))
|
|
net_load.register_buffer("buffer_var1", var)
|
|
net_load.load_dict(net.state_dict())
|
|
|
|
self.assert_var_base_equal(net_load.buffer_var1, var1)
|
|
|
|
def assert_var_base_equal(self, var1, var2):
|
|
np.testing.assert_array_equal(var1.numpy(), var2.numpy())
|
|
|
|
|
|
class TestStateDictHook(unittest.TestCase):
|
|
def test_state_dict_pre_hook(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
hook_calls = []
|
|
|
|
def state_dict_pre_hook(layer, prefix, keep_vars):
|
|
hook_calls.append((layer, prefix, keep_vars))
|
|
|
|
hook_remove_helper = layer.register_state_dict_pre_hook(
|
|
state_dict_pre_hook
|
|
)
|
|
state_dict = layer.state_dict(prefix="prefix.", keep_vars=False)
|
|
self.assertIn("prefix.weight", state_dict)
|
|
self.assertEqual(hook_calls, [(layer, "prefix.", False)])
|
|
hook_remove_helper.remove()
|
|
self.assertNotIn(
|
|
hook_remove_helper._hook_id, layer._state_dict_pre_hooks
|
|
)
|
|
|
|
def test_state_dict_post_hook(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
hook_calls = []
|
|
|
|
def state_dict_post_hook(destination):
|
|
hook_calls.append(destination)
|
|
destination["post_hook_weight"] = destination.pop("weight")
|
|
|
|
hook_remove_helper = layer.register_state_dict_post_hook(
|
|
state_dict_post_hook
|
|
)
|
|
state_dict = layer.state_dict()
|
|
self.assertIn("post_hook_weight", state_dict)
|
|
self.assertNotIn("weight", state_dict)
|
|
self.assertEqual(hook_calls, [state_dict])
|
|
hook_remove_helper.remove()
|
|
self.assertNotIn(
|
|
hook_remove_helper._hook_id, layer._state_dict_hooks
|
|
)
|
|
|
|
def test_state_dict_post_hook_with_torch_signature(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
child = paddle.nn.Layer()
|
|
child_parameter = child.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
child.register_parameter("weight", child_parameter)
|
|
layer.add_sublayer("child", child)
|
|
|
|
hook_calls = []
|
|
child_hook_calls = []
|
|
|
|
def state_dict_post_hook(
|
|
module, destination, prefix, local_metadata
|
|
):
|
|
hook_calls.append((module, destination, prefix, local_metadata))
|
|
destination[prefix + "post_hook_weight"] = destination.pop(
|
|
prefix + "weight"
|
|
)
|
|
|
|
def child_state_dict_post_hook(
|
|
module, destination, prefix, local_metadata
|
|
):
|
|
child_hook_calls.append(
|
|
(module, destination, prefix, local_metadata)
|
|
)
|
|
destination[prefix + "post_hook_weight"] = destination.pop(
|
|
prefix + "weight"
|
|
)
|
|
|
|
hook_remove_helper = layer.register_state_dict_post_hook(
|
|
state_dict_post_hook
|
|
)
|
|
child_hook_remove_helper = child.register_state_dict_post_hook(
|
|
child_state_dict_post_hook
|
|
)
|
|
|
|
state_dict = layer.state_dict(prefix="prefix.", keep_vars=False)
|
|
|
|
self.assertIn("prefix.post_hook_weight", state_dict)
|
|
self.assertIn("prefix.child.post_hook_weight", state_dict)
|
|
self.assertNotIn("prefix.weight", state_dict)
|
|
self.assertNotIn("prefix.child.weight", state_dict)
|
|
self.assertEqual(hook_calls, [(layer, state_dict, "prefix.", {})])
|
|
self.assertEqual(
|
|
child_hook_calls,
|
|
[(child, state_dict, "prefix.child.", {})],
|
|
)
|
|
|
|
hook_remove_helper.remove()
|
|
child_hook_remove_helper.remove()
|
|
self.assertNotIn(
|
|
hook_remove_helper._hook_id, layer._state_dict_hooks
|
|
)
|
|
self.assertNotIn(
|
|
child_hook_remove_helper._hook_id, child._state_dict_hooks
|
|
)
|
|
|
|
def test_state_dict_post_hook_return_with_torch_signature(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
def state_dict_post_hook(
|
|
module, destination, prefix, local_metadata
|
|
):
|
|
return {"replacement": destination[prefix + "weight"]}
|
|
|
|
layer.register_state_dict_post_hook(state_dict_post_hook)
|
|
state_dict = layer.state_dict()
|
|
|
|
self.assertEqual(list(state_dict.keys()), ["replacement"])
|
|
|
|
def test_load_state_dict_hooks(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
child = paddle.nn.Layer()
|
|
child_parameter = child.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
child.register_parameter("weight", child_parameter)
|
|
layer.add_sublayer("child", child)
|
|
|
|
pre_hook_calls = []
|
|
post_hook_calls = []
|
|
child_pre_hook_calls = []
|
|
child_post_hook_calls = []
|
|
|
|
def load_state_dict_pre_hook(
|
|
layer,
|
|
state_dict,
|
|
prefix,
|
|
local_metadata,
|
|
strict,
|
|
missing_keys,
|
|
unexpected_keys,
|
|
error_msgs,
|
|
):
|
|
pre_hook_calls.append((layer, prefix, local_metadata, strict))
|
|
|
|
def child_load_state_dict_pre_hook(
|
|
layer,
|
|
state_dict,
|
|
prefix,
|
|
local_metadata,
|
|
strict,
|
|
missing_keys,
|
|
unexpected_keys,
|
|
error_msgs,
|
|
):
|
|
child_pre_hook_calls.append(
|
|
(layer, prefix, local_metadata, strict)
|
|
)
|
|
|
|
def load_state_dict_post_hook(layer, incompatible_keys):
|
|
post_hook_calls.append((layer, incompatible_keys.missing_keys))
|
|
|
|
def child_load_state_dict_post_hook(layer, incompatible_keys):
|
|
child_post_hook_calls.append(
|
|
(layer, incompatible_keys.missing_keys)
|
|
)
|
|
|
|
pre_hook = layer.register_load_state_dict_pre_hook(
|
|
load_state_dict_pre_hook
|
|
)
|
|
post_hook = layer.register_load_state_dict_post_hook(
|
|
load_state_dict_post_hook
|
|
)
|
|
child_pre_hook = child.register_load_state_dict_pre_hook(
|
|
child_load_state_dict_pre_hook
|
|
)
|
|
child_post_hook = child.register_load_state_dict_post_hook(
|
|
child_load_state_dict_post_hook
|
|
)
|
|
|
|
incompatible_keys = layer.load_state_dict(
|
|
{
|
|
"weight": paddle.ones_like(parameter),
|
|
"child.weight": paddle.ones_like(child_parameter),
|
|
},
|
|
strict=True,
|
|
)
|
|
self.assertEqual(incompatible_keys.missing_keys, [])
|
|
self.assertEqual(incompatible_keys.unexpected_keys, [])
|
|
self.assertEqual(pre_hook_calls, [(layer, "", {}, True)])
|
|
self.assertEqual(
|
|
child_pre_hook_calls, [(child, "child.", {}, True)]
|
|
)
|
|
self.assertEqual(post_hook_calls, [(layer, [])])
|
|
self.assertEqual(child_post_hook_calls, [(child, [])])
|
|
|
|
pre_hook.remove()
|
|
post_hook.remove()
|
|
child_pre_hook.remove()
|
|
child_post_hook.remove()
|
|
self.assertNotIn(
|
|
pre_hook._hook_id, layer._load_state_dict_pre_hooks
|
|
)
|
|
self.assertNotIn(
|
|
post_hook._hook_id, layer._load_state_dict_post_hooks
|
|
)
|
|
self.assertNotIn(
|
|
child_pre_hook._hook_id, child._load_state_dict_pre_hooks
|
|
)
|
|
self.assertNotIn(
|
|
child_post_hook._hook_id, child._load_state_dict_post_hooks
|
|
)
|
|
|
|
def test_load_state_dict_post_hook_return(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
parameter = layer.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
layer.register_parameter("weight", parameter)
|
|
|
|
def load_state_dict_post_hook(layer, incompatible_keys):
|
|
return incompatible_keys
|
|
|
|
layer.register_load_state_dict_post_hook(load_state_dict_post_hook)
|
|
with self.assertRaisesRegex(
|
|
AssertionError,
|
|
"Hooks registered with ``register_load_state_dict_post_hook``",
|
|
):
|
|
layer.load_state_dict(
|
|
{"weight": paddle.ones_like(parameter)}, strict=True
|
|
)
|
|
|
|
def test_extra_state_state_dict(self):
|
|
class ExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self, value):
|
|
super().__init__()
|
|
self.value = value
|
|
|
|
def get_extra_state(self):
|
|
return {"value": self.value}
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state["value"]
|
|
|
|
with base.dygraph.guard():
|
|
layer = ExtraStateLayer(1)
|
|
layer.child = ExtraStateLayer(2)
|
|
|
|
state_dict = layer.state_dict(prefix="prefix.")
|
|
|
|
self.assertEqual(state_dict["prefix._extra_state"], {"value": 1})
|
|
self.assertEqual(
|
|
state_dict["prefix.child._extra_state"], {"value": 2}
|
|
)
|
|
|
|
def test_extra_state_load_state_dict(self):
|
|
class ExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self, value):
|
|
super().__init__()
|
|
self.value = value
|
|
parameter = self.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
self.register_parameter("weight", parameter)
|
|
|
|
def get_extra_state(self):
|
|
return {"value": self.value}
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state["value"]
|
|
|
|
with base.dygraph.guard():
|
|
layer = ExtraStateLayer(0)
|
|
layer.child = ExtraStateLayer(0)
|
|
|
|
incompatible_keys = layer.load_state_dict(
|
|
{
|
|
"weight": paddle.ones_like(layer.weight),
|
|
"_extra_state": {"value": 3},
|
|
"child.weight": paddle.ones_like(layer.child.weight),
|
|
"child._extra_state": {"value": 4},
|
|
},
|
|
strict=True,
|
|
)
|
|
|
|
self.assertEqual(layer.value, 3)
|
|
self.assertEqual(layer.child.value, 4)
|
|
np.testing.assert_array_equal(layer.weight.numpy(), np.ones([1]))
|
|
np.testing.assert_array_equal(
|
|
layer.child.weight.numpy(), np.ones([1])
|
|
)
|
|
self.assertEqual(incompatible_keys.missing_keys, [])
|
|
self.assertEqual(incompatible_keys.unexpected_keys, [])
|
|
|
|
def test_extra_state_strict_error(self):
|
|
class ExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
self.value = 0
|
|
|
|
def get_extra_state(self):
|
|
return {"value": self.value}
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state["value"]
|
|
|
|
with base.dygraph.guard():
|
|
layer = ExtraStateLayer()
|
|
layer.child = ExtraStateLayer()
|
|
|
|
with self.assertRaisesRegex(RuntimeError, "child._extra_state"):
|
|
layer.load_state_dict(
|
|
{"_extra_state": {"value": 1}}, strict=True
|
|
)
|
|
|
|
layer = paddle.nn.Layer()
|
|
with self.assertRaisesRegex(RuntimeError, "_extra_state"):
|
|
layer.load_state_dict(
|
|
{"_extra_state": {"value": 1}}, strict=True
|
|
)
|
|
|
|
def test_extra_state_non_dict(self):
|
|
class ExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self, value):
|
|
super().__init__()
|
|
self.value = value
|
|
|
|
def get_extra_state(self):
|
|
return self.value
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state
|
|
|
|
with base.dygraph.guard():
|
|
for state in ("value", 1, ExtraStateLayer(2)):
|
|
layer = ExtraStateLayer(state)
|
|
layer_load = ExtraStateLayer(None)
|
|
|
|
layer_load.load_state_dict(layer.state_dict())
|
|
|
|
self.assertEqual(layer_load.value, state)
|
|
|
|
def test_none_extra_state_is_not_saved(self):
|
|
class NoneExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
parameter = self.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
self.register_parameter("weight", parameter)
|
|
|
|
def get_extra_state(self):
|
|
return None
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state
|
|
|
|
with base.dygraph.guard():
|
|
state_dict = NoneExtraStateLayer().state_dict()
|
|
|
|
self.assertIn("weight", state_dict)
|
|
self.assertNotIn("_extra_state", state_dict)
|
|
for value in state_dict.values():
|
|
self.assertIsNotNone(value)
|
|
|
|
def test_extra_state_missing_method(self):
|
|
class MissingSetExtraStateLayer(paddle.nn.Layer):
|
|
def get_extra_state(self):
|
|
return {"value": 1}
|
|
|
|
class MissingGetExtraStateLayer(paddle.nn.Layer):
|
|
def set_extra_state(self, state):
|
|
self.value = state["value"]
|
|
|
|
with base.dygraph.guard():
|
|
layer = MissingSetExtraStateLayer()
|
|
with self.assertRaisesRegex(RuntimeError, "Unexpected key"):
|
|
layer.load_state_dict(layer.state_dict())
|
|
|
|
layer = MissingGetExtraStateLayer()
|
|
with self.assertRaisesRegex(RuntimeError, "Missing key"):
|
|
layer.load_state_dict(layer.state_dict())
|
|
|
|
def test_extra_state_with_amp_state_dict_hook(self):
|
|
class ExtraStateLayer(paddle.nn.Layer):
|
|
def __init__(self):
|
|
super().__init__()
|
|
parameter = self.create_parameter(
|
|
shape=[1], dtype='float32', is_bias=False
|
|
)
|
|
self.register_parameter("weight", parameter)
|
|
|
|
def get_extra_state(self):
|
|
return {"value": 1}
|
|
|
|
def set_extra_state(self, state):
|
|
self.value = state["value"]
|
|
|
|
with base.dygraph.guard():
|
|
layer = ExtraStateLayer()
|
|
layer = paddle.amp.decorate(
|
|
models=layer, level='O2', save_dtype='float64'
|
|
)
|
|
|
|
state_dict = layer.state_dict()
|
|
|
|
self.assertEqual(state_dict["_extra_state"], {"value": 1})
|
|
|
|
def test_default_extra_state(self):
|
|
with base.dygraph.guard():
|
|
layer = paddle.nn.Layer()
|
|
|
|
self.assertNotIn("_extra_state", layer.state_dict())
|
|
with self.assertRaisesRegex(RuntimeError, "get_extra_state"):
|
|
layer.get_extra_state()
|
|
with self.assertRaisesRegex(RuntimeError, "set_extra_state"):
|
|
layer.set_extra_state(None)
|
|
|
|
|
|
class BufferNetWithModification(paddle.nn.Layer):
|
|
def __init__(self, shape):
|
|
super().__init__()
|
|
|
|
self.buffer1 = paddle.zeros(shape, 'int32')
|
|
self.buffer2 = paddle.zeros(shape, 'int32')
|
|
|
|
@paddle.jit.to_static
|
|
def forward(self, x):
|
|
self.buffer1 += x
|
|
self.buffer2 = self.buffer1 + x
|
|
|
|
out = self.buffer1 + self.buffer2
|
|
|
|
return out
|
|
|
|
|
|
class TestModifiedBuffer(unittest.TestCase):
|
|
def funcsetUp(self):
|
|
self.shape = [10, 16]
|
|
|
|
def _run(self, to_static=False):
|
|
with enable_to_static_guard(to_static):
|
|
x = paddle.ones([1], 'int32')
|
|
net = BufferNetWithModification(self.shape)
|
|
out = net(x)
|
|
|
|
return out, net.buffer1, net.buffer2
|
|
|
|
def test_modified(self):
|
|
self.funcsetUp()
|
|
dy_outs = self._run(False)
|
|
st_outs = self._run(True)
|
|
|
|
for i in range(len(dy_outs)):
|
|
np.testing.assert_array_equal(
|
|
dy_outs[i].numpy(), st_outs[i].numpy()
|
|
)
|
|
|
|
|
|
class TestLayerTo(unittest.TestCase):
|
|
def funcsetUp(self):
|
|
self.linear = paddle.nn.Linear(2, 2)
|
|
self.new_grad = np.random.random([2, 2])
|
|
self.linear.weight._set_grad_ivar(paddle.to_tensor(self.new_grad))
|
|
buffer = paddle.to_tensor([0.0], dtype='float32')
|
|
self.linear.register_buffer("buf_name", buffer, persistable=True)
|
|
|
|
sublayer = paddle.nn.Conv1D(3, 2, 3)
|
|
self.linear.add_sublayer("1", sublayer)
|
|
|
|
def func_test_to_api(self):
|
|
if paddle.framework.use_pir_api():
|
|
dtype_float64 = paddle.base.core.DataType.FLOAT64
|
|
else:
|
|
dtype_float64 = paddle.base.core.VarDesc.VarType.FP64
|
|
self.linear.to(dtype='double')
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
|
|
self.linear.to()
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
for p in self.linear.parameters():
|
|
self.assertTrue(isinstance(p, paddle.base.framework.EagerParamBase))
|
|
|
|
if paddle.base.is_compiled_with_cuda():
|
|
self.linear.to(device=get_device_place())
|
|
self.assertTrue(self.linear.weight.place.is_gpu_place())
|
|
self.assertEqual(self.linear.weight.place.gpu_device_id(), 0)
|
|
self.assertTrue(self.linear.buf_name.place.is_gpu_place())
|
|
self.assertEqual(self.linear.buf_name.place.gpu_device_id(), 0)
|
|
self.assertTrue(
|
|
self.linear.weight._grad_ivar().place.is_gpu_place()
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().place.gpu_device_id(), 0
|
|
)
|
|
|
|
self.linear.to(device=get_device(True))
|
|
self.assertTrue(self.linear.weight.place.is_gpu_place())
|
|
self.assertEqual(self.linear.weight.place.gpu_device_id(), 0)
|
|
self.assertTrue(self.linear.buf_name.place.is_gpu_place())
|
|
self.assertEqual(self.linear.buf_name.place.gpu_device_id(), 0)
|
|
self.assertTrue(
|
|
self.linear.weight._grad_ivar().place.is_gpu_place()
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().place.gpu_device_id(), 0
|
|
)
|
|
for p in self.linear.parameters():
|
|
self.assertTrue(
|
|
isinstance(p, paddle.base.framework.EagerParamBase)
|
|
)
|
|
elif is_custom_device():
|
|
self.linear.to(device=get_device_place())
|
|
self.assertTrue(self.linear.weight.place.is_custom_place())
|
|
self.assertEqual(self.linear.weight.place.custom_device_id(), 0)
|
|
self.assertTrue(self.linear.buf_name.place.is_custom_place())
|
|
self.assertEqual(self.linear.buf_name.place.custom_device_id(), 0)
|
|
self.assertTrue(
|
|
self.linear.weight._grad_ivar().place.is_custom_place()
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().place.custom_device_id(), 0
|
|
)
|
|
|
|
self.linear.to(device=get_device(True))
|
|
self.assertTrue(self.linear.weight.place.is_custom_place())
|
|
self.assertEqual(self.linear.weight.place.custom_device_id(), 0)
|
|
self.assertTrue(self.linear.buf_name.place.is_custom_place())
|
|
self.assertEqual(self.linear.buf_name.place.custom_device_id(), 0)
|
|
self.assertTrue(
|
|
self.linear.weight._grad_ivar().place.is_custom_place()
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().place.custom_device_id(), 0
|
|
)
|
|
for p in self.linear.parameters():
|
|
self.assertTrue(
|
|
isinstance(p, paddle.base.framework.EagerParamBase)
|
|
)
|
|
|
|
self.linear.to(device=paddle.CPUPlace())
|
|
self.assertTrue(self.linear.weight.place.is_cpu_place())
|
|
self.assertTrue(self.linear.buf_name.place.is_cpu_place())
|
|
self.assertTrue(self.linear.weight._grad_ivar().place.is_cpu_place())
|
|
|
|
self.linear.to(device='cpu')
|
|
self.assertTrue(self.linear.weight.place.is_cpu_place())
|
|
self.assertTrue(self.linear.buf_name.place.is_cpu_place())
|
|
self.assertTrue(self.linear.weight._grad_ivar().place.is_cpu_place())
|
|
|
|
self.assertRaises(ValueError, self.linear.to, device=1)
|
|
|
|
self.assertRaises(TypeError, self.linear.to, blocking=1)
|
|
self.assertRaises(TypeError, self.linear.to, non_blocking=0)
|
|
|
|
def func_test_to_api_paddle_dtype(self):
|
|
if paddle.framework.use_pir_api():
|
|
dtype_float64 = paddle.base.core.DataType.FLOAT64
|
|
else:
|
|
dtype_float64 = paddle.base.core.VarDesc.VarType.FP64
|
|
|
|
self.linear.to(dtype=paddle.float64)
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
|
|
self.linear.to()
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
for p in self.linear.parameters():
|
|
self.assertTrue(isinstance(p, paddle.base.framework.EagerParamBase))
|
|
|
|
def func_test_to_api_numpy_dtype(self):
|
|
if paddle.framework.use_pir_api():
|
|
dtype_float64 = paddle.base.core.DataType.FLOAT64
|
|
else:
|
|
dtype_float64 = paddle.base.core.VarDesc.VarType.FP64
|
|
|
|
self.linear.to(dtype=np.float64)
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
|
|
self.linear.to()
|
|
self.assertEqual(self.linear.weight.dtype, dtype_float64)
|
|
self.assertEqual(self.linear.buf_name.dtype, dtype_float64)
|
|
np.testing.assert_allclose(
|
|
self.linear.weight.grad.numpy(), self.new_grad, rtol=1e-05
|
|
)
|
|
self.assertEqual(
|
|
self.linear.weight._grad_ivar().dtype,
|
|
dtype_float64,
|
|
)
|
|
for p in self.linear.parameters():
|
|
self.assertTrue(isinstance(p, paddle.base.framework.EagerParamBase))
|
|
|
|
def func_test_to_api_none_buffer(self):
|
|
model = paddle.nn.Linear(2, 4)
|
|
buffer = None
|
|
model.register_buffer("buf_name", buffer, persistable=True)
|
|
model.to(dtype='float64')
|
|
self.assertIsNone(model._buffers['buf_name'])
|
|
|
|
def test_main(self):
|
|
self.funcsetUp()
|
|
self.func_test_to_api()
|
|
self.func_test_to_api_paddle_dtype()
|
|
self.func_test_to_api_numpy_dtype()
|
|
self.func_test_to_api_none_buffer()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|