def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj,
torch.storage.TypedStorage) or torch.is_storage(obj):
if isinstance(obj, torch.storage.TypedStorage):
# TODO: Once we decide to break serialization FC, this case
# can be deleted
storage = obj._storage
storage_type_str = obj.pickle_storage_type()
storage_type = getattr(torch, storage_type_str)
storage_numel = obj.size()
else:
storage = obj
storage_type = normalize_storage_type(type(obj))
storage_numel = storage.nbytes()
storage = cast(Storage, storage)
storage_key = id_map.setdefault(storage._cdata, str(len(id_map)))
location = location_tag(storage)
serialized_storages[storage_key] = storage
return ('storage', storage_type, storage_key, location,
storage_numel)
return None
def __setitem__(self, idx, value):
if not isinstance(idx, (int, slice)):
raise RuntimeError(f"can't index a {type(self)} with {type(idx)}")
if torch.is_storage(value):
raise RuntimeError(
f'cannot set item with value type {type(value)}')
if self.dtype in [
torch.quint8, torch.quint4x2, torch.quint2x4, torch.qint32,
torch.qint8
]:
interpret_dtypes = {
torch.quint8: torch.uint8,
torch.quint4x2: torch.uint8,
torch.quint2x4: torch.uint8,
torch.qint32: torch.int32,
torch.qint8: torch.int8
}
tmp_dtype = interpret_dtypes[self.dtype]
tmp_tensor = torch.tensor(
[], dtype=tmp_dtype, device=self.device).set_(
_TypedStorage(wrap_storage=self._storage, dtype=tmp_dtype))
else:
tmp_tensor = torch.tensor([], dtype=self.dtype,
device=self.device).set_(self)
tmp_tensor[idx] = value
def _persistent_id(self, obj):
if torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
obj_key = str(obj._cdata)
location = location_tag(obj)
name = f".data/{obj_key}.storage"
if name not in self.serialized_storages:
# check to see if storage was previously serialized
serialized_files = self.zip_file.get_all_written_records()
if name not in serialized_files:
if obj.device.type != "cpu":
obj = obj.cpu()
num_bytes = obj.size() * obj.element_size()
self.zip_file.write_record(name, obj.data_ptr(), num_bytes)
self.serialized_storages.add(name)
return ("storage", storage_type, obj_key, location, obj.size())
if hasattr(obj, "__reduce_package__"):
if self.serialized_reduces.get(id(obj)) is None:
self.serialized_reduces[id(obj)] = ("reduce_package", id(obj), *obj.__reduce_package__(self))
return self.serialized_reduces[id(obj)]
return None
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_file = inspect.getsourcefile(obj)
source = inspect.getsource(obj)
except: # saving the source is optional, so we can ignore any errors
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return ('module', obj, source_file, source)
elif torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
root, offset = obj._root_storage()
root_key = str(root._cdata)
location = location_tag(obj)
serialized_storages[root_key] = root
is_view = obj._cdata != root._cdata
if is_view:
view_metadata = (str(obj._cdata), offset, obj.size())
else:
view_metadata = None
return ('storage', storage_type, root_key, location, root.size(),
view_metadata)
return None
def _persistent_id(self, obj):
if torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
location = location_tag(obj)
# serialize storage if not already written
storage_present = self.storage_context.has_storage(obj)
storage_id = self.storage_context.get_or_add_storage(obj)
if not storage_present:
if obj.device.type != "cpu":
obj = obj.cpu()
num_bytes = obj.size() * obj.element_size()
self.zip_file.write_record(f".data/{storage_id}.storage",
obj.data_ptr(), num_bytes)
return ("storage", storage_type, storage_id, location, obj.size())
if hasattr(obj, "__reduce_package__"):
if _gate_torchscript_serialization and isinstance(
obj, torch.jit.RecursiveScriptModule):
raise Exception(
"Serializing ScriptModules directly into a package is a beta feature. "
"To use, set global "
"`torch.package.package_exporter._gate_torchscript_serialization` to `False`."
)
if self.serialized_reduces.get(id(obj)) is None:
self.serialized_reduces[id(obj)] = (
"reduce_package",
id(obj),
*obj.__reduce_package__(self),
)
return self.serialized_reduces[id(obj)]
return None
def _persistent_id(self, obj):
if torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
obj_key = str(obj._cdata)
location = location_tag(obj)
name = f".data/{obj_key}.storage"
if name not in self.serialized_storages:
# check to see if storage was previously serialized
serialized_files = self.zip_file.get_all_written_records()
if name not in serialized_files:
if obj.device.type != "cpu":
obj = obj.cpu()
num_bytes = obj.size() * obj.element_size()
self.zip_file.write_record(name, obj.data_ptr(), num_bytes)
self.serialized_storages.add(name)
return ("storage", storage_type, obj_key, location, obj.size())
if hasattr(obj, "__reduce_package__"):
if _gate_torchscript_serialization and isinstance(
obj, torch.jit.RecursiveScriptModule):
raise Exception(
"Serializing ScriptModules directly into a package is a beta feature. "
"To use, set global "
"`torch.package.package_exporter._gate_torchscript_serialization` to `False`."
)
if self.serialized_reduces.get(id(obj)) is None:
self.serialized_reduces[id(obj)] = (
"reduce_package", id(obj), *obj.__reduce_package__(self))
return self.serialized_reduces[id(obj)]
return None
def persistent_id(obj):
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_lines, _, source_file = get_source_lines_and_file(obj)
source = ''.join(source_lines)
except Exception: # saving the source is optional, so we can ignore any errors
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return ('module', obj, source_file, source)
elif torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
# Offset is always 0, but we keep it for backwards compatibility
# with the old serialization format (which supported storage views)
offset = 0
obj_key = str(obj._cdata)
location = location_tag(obj)
serialized_storages[obj_key] = obj
is_view = obj._cdata != obj._cdata
if is_view:
view_metadata = (str(obj._cdata), offset, obj.size())
else:
view_metadata = None
return ('storage', storage_type, obj_key, location, obj.size(),
view_metadata)
return None
def persistent_id(obj):
if torch.is_storage(obj) or isinstance(obj,
torch.storage.TypedStorage):
if isinstance(obj, torch.storage.TypedStorage):
# TODO: Once we decide to break serialization FC, we can
# remove this case
storage = obj._storage
dtype = obj.dtype
else:
storage = obj
dtype = torch.uint8
serialized_storages.append(obj)
serialized_dtypes.append(dtype)
return ("storage", len(serialized_storages) - 1)
if hasattr(obj, "__reduce_deploy__"):
if _serialized_reduces.get(id(obj)) is None:
_serialized_reduces[id(obj)] = (
"reduce_deploy",
id(obj),
*obj.__reduce_deploy__(importers),
)
return _serialized_reduces[id(obj)]
return None
def persistent_id(obj):
if torch.is_tensor(obj):
serialized_tensors[obj._cdata] = obj
return str(obj._cdata)
elif torch.is_storage(obj):
serialized_storages[obj._cdata] = obj
return str(obj._cdata)
return None
def storage_ptr(obj):
if torch.is_tensor(obj):
return obj.storage().data_ptr()
elif torch.is_storage(obj):
return obj.data_ptr()
elif is_new_api() and isinstance(obj, typed_storage_class):
return obj._storage.data_ptr()
else:
assert False, f'type {type(obj)} is not supported'
def test1():
"""
# check whether an object in Python is a tensor
:return: bool
"""
x = [0, 1, 2]
print("whether x is tensor :",
torch.is_tensor(x)) # check whether is tensor -> False
print("whether x is storage :",
torch.is_storage(x)) # check whether is stored -> False
y = torch.randn(3, 2) # shape=(3, 2) / torch.zeros(3, 2)
print("whether y is tensor :",
torch.is_tensor(y)) # check whether is tensor -> True
print("whether y is storage :",
torch.is_storage(y)) # check whether is stored -> False
print("the total number of elements in the input Tensor is : {}".format(
torch.numel(y)))
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if torch.is_storage(obj):
serialized_storages.append(obj)
serialized_dtypes.append(obj.dtype)
return ('storage', len(serialized_storages) - 1)
return None
def _persistent_id(self, obj):
if torch.is_storage(obj) or isinstance(obj,
torch.storage._TypedStorage):
if isinstance(obj, torch.storage._TypedStorage):
# TODO: Once we decide to break serialization FC, we can
# remove this case
untyped_storage = obj._storage
storage_type_str = obj.pickle_storage_type()
storage_type = getattr(torch, storage_type_str)
dtype = obj.dtype
storage_numel = obj.size()
elif isinstance(obj, torch._UntypedStorage):
untyped_storage = obj
storage_type = normalize_storage_type(type(storage))
dtype = torch.uint8
storage_numel = storage.nbytes()
else:
raise RuntimeError(f"storage type not recognized: {type(obj)}")
storage: Storage = cast(Storage, untyped_storage)
location = location_tag(storage)
# serialize storage if not already written
storage_present = self.storage_context.has_storage(storage)
storage_id = self.storage_context.get_or_add_storage(storage)
if not storage_present:
if storage.device.type != "cpu":
storage = storage.cpu()
num_bytes = storage.nbytes()
self.zip_file.write_record(f".data/{storage_id}.storage",
storage.data_ptr(), num_bytes)
return ("storage", storage_type, storage_id, location,
storage_numel)
if hasattr(obj, "__reduce_package__"):
if _gate_torchscript_serialization and isinstance(
obj, torch.jit.RecursiveScriptModule):
raise Exception(
"Serializing ScriptModules directly into a package is a beta feature. "
"To use, set global "
"`torch.package.package_exporter._gate_torchscript_serialization` to `False`."
)
if self.serialized_reduces.get(id(obj)) is None:
self.serialized_reduces[id(obj)] = (
"reduce_package",
id(obj),
*obj.__reduce_package__(self),
)
return self.serialized_reduces[id(obj)]
return None
def _persistent_id(self, obj):
if torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
obj_key = str(obj._cdata)
location = location_tag(obj)
self.serialized_storages[obj_key] = obj
return ("storage", storage_type, obj_key, location, obj.size())
if hasattr(obj, "__reduce_package__"):
return ("reduce_package", *obj.__reduce_package__(self))
return None
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
obj_key = str(obj._cdata)
location = location_tag(obj)
serialized_storages[obj_key] = obj
return ('storage', storage_type, obj_key, location, obj.size())
return None
def safe_call(*args, **kwargs):
args = tuple(
ffi.cast(_torch_to_cffi.get(type(arg), 'void') + '*', arg._cdata)
if torch.is_tensor(arg) or torch.is_storage(arg) else arg
for arg in args)
args = (function, ) + args
result = torch._C._safe_call(*args, **kwargs)
if isinstance(result, ffi.CData):
typeof = ffi.typeof(result)
if typeof.kind == 'pointer':
cdata = int(ffi.cast('uintptr_t', result))
cname = typeof.item.cname
if cname in _cffi_to_torch:
return _cffi_to_torch[cname](cdata=cdata)
return result
def persistent_id(obj):
if isinstance(obj, type) and issubclass(obj, nn.Container):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = inspect.getsourcefile(obj)
source = inspect.getsource(obj)
return (obj, source_file, source)
if torch.is_tensor(obj):
serialized_tensors[obj._cdata] = obj
return str(obj._cdata)
elif torch.is_storage(obj):
serialized_storages[obj._cdata] = obj
return str(obj._cdata)
return None
def safe_call(*args, **kwargs):
args = tuple(ffi.cast(_torch_to_cffi.get(type(arg), 'void') + '*', arg._cdata)
if torch.is_tensor(arg) or torch.is_storage(arg)
else arg
for arg in args)
args = (function,) + args
result = torch._C._safe_call(*args, **kwargs)
if isinstance(result, ffi.CData):
typeof = ffi.typeof(result)
if typeof.kind == 'pointer':
cdata = int(ffi.cast('uintptr_t', result))
cname = typeof.item.cname
if cname in _cffi_to_torch:
return _cffi_to_torch[cname](cdata=cdata)
return result
def to_gpu(obj, type_map={}):
if isinstance(obj, torch.Tensor):
assert obj.is_leaf
t = type_map.get(obj.type(), get_gpu_type(obj.type()))
with torch.no_grad():
res = obj.clone().type(t)
res.requires_grad = obj.requires_grad
return res
elif torch.is_storage(obj):
return obj.new().resize_(obj.size()).copy_(obj)
elif isinstance(obj, list):
return [to_gpu(o, type_map) for o in obj]
elif isinstance(obj, tuple):
return tuple(to_gpu(o, type_map) for o in obj)
else:
return deepcopy(obj)
def to_gpu(obj, type_map={}):
if torch.is_tensor(obj):
t = type_map.get(type(obj), get_gpu_type(type(obj)))
return obj.clone().type(t)
elif torch.is_storage(obj):
return obj.new().resize_(obj.size()).copy_(obj)
elif isinstance(obj, Variable):
assert obj.is_leaf
t = type_map.get(type(obj.data), get_gpu_type(type(obj.data)))
return Variable(obj.data.clone().type(t), requires_grad=obj.requires_grad)
elif isinstance(obj, list):
return [to_gpu(o, type_map) for o in obj]
elif isinstance(obj, tuple):
return tuple(to_gpu(o, type_map) for o in obj)
else:
return deepcopy(obj)
def to_gpu(obj, type_map={}):
if isinstance(obj, torch.Tensor):
assert obj.is_leaf
t = type_map.get(obj.type(), get_gpu_type(obj.type()))
with torch.no_grad():
res = obj.clone().type(t)
res.requires_grad = obj.requires_grad
return res
elif torch.is_storage(obj):
return obj.new().resize_(obj.size()).copy_(obj)
elif isinstance(obj, list):
return [to_gpu(o, type_map) for o in obj]
elif isinstance(obj, tuple):
return tuple(to_gpu(o, type_map) for o in obj)
else:
return deepcopy(obj)
def to_gpu(obj, type_map={}):
if torch.is_tensor(obj):
t = type_map.get(type(obj), get_gpu_type(type(obj)))
return obj.clone().type(t)
elif torch.is_storage(obj):
return obj.new().resize_(obj.size()).copy_(obj)
elif isinstance(obj, Variable):
assert obj.is_leaf
t = type_map.get(type(obj.data), get_gpu_type(type(obj.data)))
return Variable(obj.data.clone().type(t), requires_grad=obj.requires_grad)
elif isinstance(obj, list):
return [to_gpu(o, type_map) for o in obj]
elif isinstance(obj, tuple):
return tuple(to_gpu(o, type_map) for o in obj)
else:
return deepcopy(obj)
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj,
torch.storage._TypedStorage) or torch.is_storage(obj):
if isinstance(obj, torch.storage._TypedStorage):
# TODO: Once we decide to break serialization FC, this case
# can be deleted
storage = obj._storage
storage_dtype = obj.dtype
storage_type_str = obj.pickle_storage_type()
storage_type = getattr(torch, storage_type_str)
storage_numel = obj.size()
else:
storage = obj
storage_dtype = storage.dtype
storage_type = normalize_storage_type(type(obj))
storage_numel = storage.nbytes()
storage = cast(Storage, storage)
# If storage is allocated, ensure that any other saved storages
# pointing to the same data all have the same dtype. If storage is
# not allocated, don't perform this check
if storage.data_ptr() != 0:
if storage.data_ptr() in storage_dtypes:
if storage_dtype != storage_dtypes[storage.data_ptr()]:
raise RuntimeError(
'Cannot save multiple tensors or storages that '
'view the same data as different types')
else:
storage_dtypes[storage.data_ptr()] = storage_dtype
storage_key = id_map.setdefault(storage._cdata, str(len(id_map)))
location = location_tag(storage)
serialized_storages[storage_key] = storage
return ('storage', storage_type, storage_key, location,
storage_numel)
return None
def safe_call(*args, **kwargs):
args = tuple(
ffi.cast(_torch_to_cffi.get(arg.type(), 'void') + '*', arg._cdata)
if isinstance(arg, torch.Tensor) or torch.is_storage(arg) else arg
for arg in args)
args = (function, ) + args
result = torch._C._safe_call(*args, **kwargs)
if isinstance(result, ffi.CData):
typeof = ffi.typeof(result)
if typeof.kind == 'pointer':
cdata = int(ffi.cast('uintptr_t', result))
cname = typeof.item.cname
if cname in _cffi_to_torch:
# TODO: Maybe there is a less janky way to eval
# off of this
return eval(_cffi_to_torch[cname])(cdata=cdata)
return result
def safe_call(*args, **kwargs):
args = tuple(ffi.cast(_torch_to_cffi.get(arg.type(), 'void') + '*', arg._cdata)
if isinstance(arg, torch.Tensor) or torch.is_storage(arg)
else arg
for arg in args)
args = (function,) + args
result = torch._C._safe_call(*args, **kwargs)
if isinstance(result, ffi.CData):
typeof = ffi.typeof(result)
if typeof.kind == 'pointer':
cdata = int(ffi.cast('uintptr_t', result))
cname = typeof.item.cname
if cname in _cffi_to_torch:
# TODO: Maybe there is a less janky way to eval
# off of this
return eval(_cffi_to_torch[cname])(cdata=cdata)
return result
def persistent_id(obj: Any) -> Optional[Tuple]:
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_lines, _, source_file = get_source_lines_and_file(obj)
source = ''.join(source_lines)
except Exception: # saving the source is optional, so we can ignore any errors
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return ('module', obj, source_file, source)
elif torch.is_storage(obj):
view_metadata: Optional[Tuple[str, int, int]]
obj = cast(Storage, obj)
storage_type = normalize_storage_type(type(obj))
# Offset is always 0, but we keep it for backwards compatibility
# with the old serialization format (which supported storage views)
offset = 0
obj_key = str(obj._cdata)
location = location_tag(obj)
serialized_storages[obj_key] = obj
is_view = obj._cdata != obj._cdata
if is_view:
view_metadata = (str(obj._cdata), offset, obj.size())
else:
view_metadata = None
return ('storage',
storage_type,
obj_key,
location,
obj.size(),
view_metadata)
return None
def tensor():
print('Hello, PyTorch.tensor')
y = torch.eye(3, 4)
print(y)
print(torch.is_tensor(y))
print(torch.is_storage(y))
print(torch.numel(y))
x = [12, 23, 34, 45, 56, 67, 78]
x1 = np.array(x)
print(torch.from_numpy(x1))
print(torch.linspace(2, 10, steps=25))
print(torch.logspace(-10, 10, steps=15))
print(torch.rand(10))
print(torch.rand(4, 5))
print(torch.randn(10))
print(torch.randperm(10))
print(torch.arange(10, 40, 2))
d = torch.randn(4, 5)
print(d)
print(torch.argmin(d))
print(torch.argmin(d, dim=1))
print(torch.zeros(4, 5))
x = torch.randn(4, 5)
print(x)
print(torch.cat((x, x)))
print(torch.cat((x, x, x), 1))
a = torch.randn(4, 4)
print(a)
print(torch.chunk(a, 2))
print(torch.chunk(a, 2, 1))
print(
torch.gather(torch.tensor([[11, 12], [23, 24]]), 1,
torch.LongTensor([[0, 0], [1, 0]])))
x = torch.randn(4, 5)
print(x)
print(x.t())
print(x.transpose(1, 0))
def check_regular_serialization(loaded_list, check_list):
for idx0 in range(len(check_list)):
check_val0 = check_list[idx0]
loaded_val0 = loaded_list[idx0]
# Check that loaded values are what they should be
assert type(check_val0) == type(loaded_val0), (
f'type should be {type(check_val0)} but got {type(loaded_val0)}')
if torch.is_tensor(check_val0):
assert check_val0.device == loaded_val0.device
assert check_val0.eq(loaded_val0).all()
elif torch.is_storage(check_val0):
assert check_val0.device == loaded_val0.device
assert check_val0.tolist() == loaded_val0.tolist()
elif issubclass(type(check_val0), torch.nn.Module):
param_pairs = zip(check_val0.parameters(),
loaded_val0.parameters())
assert all([p0.device == p1.device for p0, p1 in param_pairs])
assert all([p0.eq(p1).all() for p0, p1 in param_pairs])
elif is_new_api() and isinstance(check_val0, typed_storage_class):
assert check_val0._storage.device == loaded_val0._storage.device
assert check_val0.dtype == loaded_val0.dtype
assert check_val0._storage.tolist() == loaded_val0._storage.tolist(
)
else:
assert False, f'type {type(check_val0)} not supported'
if not has_data_ptr(check_val0):
continue
# Check that storage sharing is preserved
for idx1 in range(idx0 + 1, len(check_list)):
check_val1 = check_list[idx1]
loaded_val1 = loaded_list[idx0]
if storage_ptr(check_val0) == storage_ptr(check_val1):
assert storage_ptr(loaded_val0) == storage_ptr(loaded_val1)
def to_gpu(obj, type_map={}):
if torch.is_tensor(obj):
t = type_map.get(type(obj), get_gpu_type(type(obj)))
# Workaround since torch.HalfTensor doesn't support clone()
if type(obj) == torch.HalfTensor:
return obj.new().resize_(obj.size()).copy_(obj).type(t)
return obj.clone().type(t)
elif torch.is_storage(obj):
return obj.new().resize_(obj.size()).copy_(obj)
elif isinstance(obj, Variable):
assert obj.is_leaf
t = type_map.get(type(obj.data), get_gpu_type(type(obj.data)))
o = obj.type(t).detach()
o.requires_grad = obj.requires_grad
return o
elif isinstance(obj, list):
return [to_gpu(o, type_map) for o in obj]
elif isinstance(obj, tuple):
return tuple(to_gpu(o, type_map) for o in obj)
else:
return deepcopy(obj)
def persistent_id(obj):
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_file = inspect.getsourcefile(obj)
source = inspect.getsource(obj)
except (TypeError, IOError):
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return (obj, source_file, source)
if torch.is_tensor(obj):
serialized_tensors[obj._cdata] = obj
return str(obj._cdata)
elif torch.is_storage(obj):
serialized_storages[obj._cdata] = obj
return str(obj._cdata)
return None
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if torch.is_storage(obj):
serialized_storages.append(obj)
serialized_dtypes.append(obj.dtype)
return ('storage', len(serialized_storages) - 1)
if hasattr(obj, "__reduce_deploy__"):
if _serialized_reduces.get(id(obj)) is None:
_serialized_reduces[id(obj)] = (
"reduce_deploy",
id(obj),
*obj.__reduce_deploy__(importers),
)
return _serialized_reduces[id(obj)]
return None
def persistent_id(obj):
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_file = inspect.getsourcefile(obj)
source = inspect.getsource(obj)
except Exception: # saving the source is optional, so we can ignore any errors
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return ('module', obj, source_file, source)
elif torch.is_storage(obj):
storage_type = normalize_storage_type(type(obj))
root, offset = obj._root_storage()
root_key = str(root._cdata)
location = location_tag(obj)
serialized_storages[root_key] = root
is_view = obj._cdata != root._cdata
if is_view:
view_metadata = (str(obj._cdata), offset, obj.size())
else:
view_metadata = None
return ('storage',
storage_type,
root_key,
location,
root.size(),
view_metadata)
return None
def persistent_id(obj: Any) -> Optional[Tuple]:
# FIXME: the docs say that persistent_id should only return a string
# but torch store returns tuples. This works only in the binary protocol
# see
# https://docs.python.org/2/library/pickle.html#pickling-and-unpickling-external-objects
# https://github.com/python/cpython/blob/master/Lib/pickle.py#L527-L537
if isinstance(obj, type) and issubclass(obj, nn.Module):
if obj in serialized_container_types:
return None
serialized_container_types[obj] = True
source_file = source = None
try:
source_lines, _, source_file = get_source_lines_and_file(obj)
source = ''.join(source_lines)
except Exception: # saving the source is optional, so we can ignore any errors
warnings.warn("Couldn't retrieve source code for container of "
"type " + obj.__name__ + ". It won't be checked "
"for correctness upon loading.")
return ('module', obj, source_file, source)
if isinstance(obj,
torch.storage._TypedStorage) or torch.is_storage(obj):
if isinstance(obj, torch.storage._TypedStorage):
# TODO: Once we decide to break serialization FC, this case
# can be deleted
storage = obj._storage
storage_dtype = obj.dtype
storage_type_str = obj.pickle_storage_type()
storage_type = getattr(torch, storage_type_str)
dtype = obj.dtype
storage_numel = obj.size()
else:
storage = obj
storage_dtype = storage.dtype
storage_type = normalize_storage_type(type(obj))
dtype = torch.uint8
storage_numel = cast(Storage, storage).nbytes()
# If storage is allocated, ensure that any other saved storages
# pointing to the same data all have the same dtype. If storage is
# not allocated, don't perform this check
if storage.data_ptr() != 0:
if storage.data_ptr() in storage_dtypes:
if storage_dtype != storage_dtypes[storage.data_ptr()]:
raise RuntimeError(
'Cannot save multiple tensors or storages that '
'view the same data as different types')
else:
storage_dtypes[storage.data_ptr()] = storage_dtype
view_metadata: Optional[Tuple[str, int, int]]
storage = cast(Storage, storage)
# Offset is always 0, but we keep it for backwards compatibility
# with the old serialization format (which supported storage views)
offset = 0
storage_key = str(storage._cdata)
location = location_tag(storage)
# TODO: There's an issue here with FC. It might be impossible to
# solve, but it's worth noting. Imagine we save a list `[storage,
# tensor]`, where `tensor.storage()` is the same as `storage`, and
# `tensor.element_size() > 1`. Let's say that `tensor.dtype ==
# torch.float`. The storage will be serialized with element size
# of 1, since we're choosing to serialize the first occurance of
# a duplicate storage. Since this legacy serialization format saves
# the numel of the storage, rather than nbytes directly, we'll be
# effectively saving nbytes in this case. We'll be able to load it
# and the tensor back up with no problems in _this_ and future
# versions of pytorch, but in older versions, here's the problem:
# the storage will be loaded up as a _UntypedStorage, and then the
# FloatTensor will loaded and the _UntypedStorage will be assigned to
# it. Since the storage dtype does not match the tensor dtype, this
# will cause an error. If we reverse the list, like `[tensor,
# storage]`, then we will save the `tensor.storage()` as a faked
# `FloatStorage`, and the saved size will be the correct
# dtype-specific numel count that old versions expect. `tensor`
# will be able to load up properly in old versions, pointing to
# a FloatStorage. However, `storage` is still being translated to
# a _UntypedStorage, and it will try to resolve to the same
# FloatStorage that `tensor` contains. This will also cause an
# error. It doesn't seem like there's any way around this.
# Probably, we just cannot maintain FC for the legacy format if the
# saved list contains both a tensor and a storage that point to the
# same data. We should still be able to maintain FC for lists of
# just tensors, as long as all views share the same dtype as the
# tensor they are viewing.
if storage_key not in serialized_storages:
serialized_storages[storage_key] = (storage, dtype)
is_view = storage._cdata != storage._cdata
if is_view:
view_metadata = (str(storage._cdata), offset, storage.nbytes())
else:
view_metadata = None
res = ('storage', storage_type, storage_key, location,
storage_numel, view_metadata)
return res
return None
def basic_operation():
# REF [site] >>
# https://pytorch.org/docs/stable/tensors.html
# https://pytorch.org/docs/stable/tensor_attributes.html
x = torch.empty(5, 3)
print('x =', x)
print('x.shape = {}, x.dtype = {}.'.format(x.shape, x.dtype))
#print('x =', x.data)
x = torch.rand(2, 3)
print('x =', x)
x = torch.randn(2, 3)
print('x =', x)
x = torch.randn(2, 3)
print('x =', x)
x = torch.randperm(5)
print('x =', x)
x = torch.FloatTensor(10, 12, 3, 3)
print('x =', x.size())
print('x =', x.size()[:])
#--------------------
y = torch.zeros(2, 3)
print('y =', y)
y = torch.ones(2, 3)
print('y =', y)
y = torch.arange(0, 3, step=0.5)
print('y =', y)
x = torch.tensor(1, dtype=torch.int32)
#x = torch.tensor(1, dtype=torch.int32, device='cuda:1')
print('x =', x)
x = torch.tensor([5.5, 3])
print('x =', x)
x = x.new_ones(5, 3, dtype=torch.double) # new_* methods take in sizes.
print('x =', x)
x = torch.randn_like(x, dtype=torch.float) # Override dtype.
print('x =', x)
#--------------------
y = torch.rand(5, 3)
print('x + y =', x + y)
print('x + y =', torch.add(x, y))
result = torch.empty(5, 3)
torch.add(x, y, out=result)
print('x + y =', result)
#--------------------
# Any operation that mutates a tensor in-place is post-fixed with an _.
# For example: x.copy_(y), x.t_(), will change x.
y.add_(x) # In-place.
print('y =', y)
#--------------------
# You can use standard NumPy-like indexing with all bells and whistles!
print(x[:, 1])
#--------------------
# If you have a one element tensor, use .item() to get the value as a Python number.
x = torch.randn(1)
print('x =', x)
print('x.item() =', x.item())
#--------------------
x = torch.randn(2, 2)
print('x.is_cuda =', x.is_cuda)
print('x.is_complex() =', x.is_complex())
print('x.is_contiguous() =', x.is_contiguous())
print('x.is_distributed() =', x.is_distributed())
print('x.is_floating_point() =', x.is_floating_point())
print('x.is_pinned() =', x.is_pinned())
print('x.is_quantized =', x.is_quantized)
print('x.is_shared() =', x.is_shared())
print('x.is_signed() =', x.is_signed())
print('x.is_sparse =', x.is_sparse)
print('x.contiguous() =', x.contiguous())
print('x.storage() =', x.storage())
#--------------------
x = torch.randn(2, 2)
print('torch.is_tensor(x) =', torch.is_tensor(x))
print('torch.is_storage(x) =', torch.is_storage(x))
print('torch.is_complex(x) =', torch.is_complex(x))
print('torch.is_floating_point(x) =', torch.is_floating_point(x))
# Sets the default floating point dtype to d.
# This type will be used as default floating point type for type inference in torch.tensor().
torch.set_default_dtype(torch.float32)
print('torch.get_default_dtype() =', torch.get_default_dtype())
# Sets the default torch.Tensor type to floating point tensor type.
# This type will also be used as default floating point type for type inference in torch.tensor().
torch.set_default_tensor_type(torch.FloatTensor)
#--------------------
# REF [site] >> https://pytorch.org/docs/stable/tensor_view.html
# View tensor shares the same underlying data with its base tensor.
# Supporting View avoids explicit data copy, thus allows us to do fast and memory efficient reshaping, slicing and element-wise operations.
# If you want to resize/reshape tensor, you can use torch.view.
x = torch.randn(4, 4)
y = x.view(16)
z = x.view(-1, 8) # The size -1 is inferred from other dimensions.
print('x.size() = {}, y.size() = {}, z.size() = {}.'.format(
x.size(), y.size(), z.size()))
t = torch.rand(4, 4)
b = t.view(2, 8)
print('t.storage().data_ptr() == b.storage().data_ptr()?',
t.storage().data_ptr() == b.storage().data_ptr())
def has_data_ptr(obj):
return torch.is_tensor(obj) or torch.is_storage(obj) or (
is_new_api() and isinstance(obj, typed_storage_class))