def decorate_fwd(*args, **kwargs):
if cast_inputs is None:
args[0]._fwd_used_autocast = torch.is_autocast_enabled()
return fwd(*args, **kwargs)
else:
autocast_context = torch.is_autocast_enabled()
args[0]._fwd_used_autocast = False
if autocast_context:
with autocast(enabled=False):
return fwd(*_cast(args, cast_inputs), **_cast(kwargs, cast_inputs))
else:
return fwd(*args, **kwargs)
def forward(self,
graph: DGLGraph,
node_feats: Dict[str, Tensor],
edge_feats: Optional[Dict[str, Tensor]] = None,
basis: Optional[Dict[str, Tensor]] = None):
# Compute bases in case they weren't precomputed as part of the data loading
basis = basis or get_basis(graph.edata['rel_pos'],
max_degree=self.max_degree,
compute_gradients=False,
use_pad_trick=self.tensor_cores
and not self.low_memory,
amp=torch.is_autocast_enabled())
# Add fused bases (per output degree, per input degree, and fully fused) to the dict
basis = update_basis_with_fused(
basis,
self.max_degree,
use_pad_trick=self.tensor_cores and not self.low_memory,
fully_fused=self.fuse_level == ConvSE3FuseLevel.FULL)
edge_feats = get_populated_edge_features(graph.edata['rel_pos'],
edge_feats)
node_feats = self.graph_modules(node_feats,
edge_feats,
graph=graph,
basis=basis)
if self.pooling is not None:
return self.pooling_module(node_feats, graph=graph)
if self.return_type is not None:
return node_feats[str(self.return_type)]
return node_feats
def criterion_parallel_apply(modules, inputs, targets, devices, kwargs_tup=None):
if kwargs_tup is None:
kwargs_tup = ({},) * len(modules)
lock = Lock()
results = {}
grad_enabled, autocast_enabled = torch.is_grad_enabled(), torch.is_autocast_enabled()
def _worker(i, module, input, target, kwargs, device):
if not isinstance(input, (list, tuple)):
input = (input,)
if not isinstance(target, (list, tuple)):
target = (target,)
with torch.set_grad_enabled(grad_enabled), torch.cuda.device(device), autocast(enabled=autocast_enabled):
output = module(*(input + target), **kwargs)
with lock:
results[i] = output
threads = [Thread(target=_worker, args=(i, module, input, target, kwargs, device)) for i, (module, input, target, kwargs, device) in enumerate(zip(modules, inputs, targets, kwargs_tup, devices))]
for thread in threads:
thread.start()
for thread in threads:
thread.join()
outputs = []
for i in range(len(inputs)):
output = results[i]
outputs.append(output)
return outputs
def forward(ctx, run_function, preserve_rng_state, *args):
check_backward_validity(args)
ctx.run_function = run_function
ctx.preserve_rng_state = preserve_rng_state
ctx.had_autocast_in_fwd = torch.is_autocast_enabled()
if preserve_rng_state:
ctx.fwd_cpu_state = torch.get_rng_state()
# Don't eagerly initialize the cuda context by accident.
# (If the user intends that the context is initialized later, within their
# run_function, we SHOULD actually stash the cuda state here. Unfortunately,
# we have no way to anticipate this will happen before we run the function.)
ctx.had_cuda_in_fwd = False
if torch.cuda._initialized:
ctx.had_cuda_in_fwd = True
ctx.fwd_gpu_devices, ctx.fwd_gpu_states = get_device_states(
*args)
# Save non-tensor inputs in ctx, keep a placeholder None for tensors
# to be filled out during the backward.
ctx.inputs = []
ctx.tensor_indices = []
tensor_inputs = []
for i, arg in enumerate(args):
if torch.is_tensor(arg):
tensor_inputs.append(arg)
ctx.tensor_indices.append(i)
ctx.inputs.append(None)
else:
ctx.inputs.append(arg)
ctx.save_for_backward(*tensor_inputs)
with torch.no_grad():
outputs = run_function(*args)
return outputs
def __enter__(self):
if torch._jit_internal.is_scripting():
assert self.fast_dtype is not None
return self
self.prev_cache_enabled = torch.is_autocast_cache_enabled()
if self.device == 'cpu':
self.prev = torch.is_autocast_cpu_enabled()
self.prev_fastdtype = torch.get_autocast_cpu_dtype()
torch.set_autocast_cpu_enabled(self._enabled)
torch.set_autocast_cpu_dtype(
self.fast_dtype) # type: ignore[arg-type]
torch.autocast_increment_nesting()
elif self.device == 'xpu':
self.prev = torch.xpu.is_autocast_xpu_enabled(
) # type: ignore[attr-defined]
self.prev_fastdtype = torch.xpu.get_autocast_xpu_dtype(
) # type: ignore[attr-defined]
torch.xpu.set_autocast_xpu_enabled(
self._enabled) # type: ignore[attr-defined]
torch.xpu.set_autocast_xpu_dtype(
self.fast_dtype) # type: ignore[attr-defined]
torch.autocast_increment_nesting()
else:
self.prev = torch.is_autocast_enabled()
self.prev_fastdtype = torch.get_autocast_gpu_dtype()
torch.set_autocast_gpu_dtype(
self.fast_dtype) # type: ignore[arg-type]
torch.set_autocast_enabled(self._enabled)
torch.autocast_increment_nesting()
torch.set_autocast_cache_enabled(self._cache_enabled)
def forward(ctx, run_function, preserve_rng_state, *args):
check_backward_validity(args)
ctx.run_function = run_function
ctx.preserve_rng_state = preserve_rng_state
ctx.has_autocast_in_fwd = torch.is_autocast_enabled()
if preserve_rng_state:
ctx.fwd_cpu_state = torch.get_rng_state()
ctx.had_cuda_in_fwd = False
if torch.cuda._initialized:
ctx.had_cuda_in_fwd = True
ctx.device_states = get_device_states(*args)
def replace_tensors(arg):
if torch.is_tensor(arg):
return TensorPlaceholder(None)
return arg
tensor_inputs = []
ctx.inputs = recursive_apply(replace_tensors, args)
for input_arg, all_arg in zip(recursive_walk(ctx.inputs),
recursive_walk(args)):
if isinstance(input_arg, TensorPlaceholder):
assert torch.is_tensor(all_arg)
assert input_arg.tensor_index is None
tensor_inputs.append(all_arg)
input_arg.tensor_index = len(tensor_inputs) - 1
ctx.save_for_backward(*tensor_inputs)
with torch.no_grad():
outputs = run_function(*args)
for output in recursive_walk(outputs):
if torch.is_tensor(output):
output.requires_grad_(True)
return outputs
def batch_norm(g, input, weight, bias, running_mean, running_var, training,
momentum, eps, cudnn_enabled):
if torch.is_autocast_enabled() and \
not args_have_same_dtype([input, weight, bias, running_mean, running_var]) and \
sym_help._export_onnx_opset_version < 15:
return sym_help._onnx_opset_unsupported_detailed(
"BatchNormalization", 14, 15,
"All input tensors must have the same `dtype`."
" Turn off Autocast or export using opset version 15.")
sym_help.check_training_mode(training, "batch_norm")
weight, bias, running_mean, running_var = sym_help._batchnorm_helper(
g, input, weight, bias, running_mean, running_var)
out = g.op("BatchNormalization",
input,
weight,
bias,
running_mean,
running_var,
epsilon_f=eps,
momentum_f=1 - momentum,
training_mode_i=0 if not training else 1,
outputs=1 if not training else 3)
if not training:
return out
else:
res, new_running_mean, new_running_var = out
new_running_mean.setType(running_mean.type())
new_running_var.setType(running_var.type())
return res
def fn(x):
if torch.is_autocast_enabled():
y = True
else:
y = False
with torch.cuda.amp.autocast(enabled=True):
z = x.relu()
return y, z
def forward(self, *args, **kwargs):
if torch.is_autocast_enabled():
with torch.cuda.amp.autocast(enabled=False):
result = self.module(*tensor_to(args, torch.float32),
**tensor_to(kwargs, torch.float32))
else:
result = self.module(*args, **kwargs)
return result
def fix_inf(self, x):
if not self.detect_inf:
return x
# to enable fp16 training
is_fp16 = x.dtype == torch.float16 or torch.is_autocast_enabled()
if is_fp16 and torch.isinf(x).any():
clamp_value = torch.finfo(torch.float16).max - 1000
x = torch.clamp(x, min=-clamp_value, max=clamp_value)
return x
def on_batch_start(self, runner: IRunner) -> None:
"""On batch start event
Args:
runner: current runner
"""
self.prev_autocast_state = torch.is_autocast_enabled()
torch.set_autocast_enabled(True)
torch.autocast_increment_nesting()
def _get_autocast_kwargs():
gpu_autocast_kwargs = {"enabled": torch.is_autocast_enabled(),
"dtype": torch.get_autocast_gpu_dtype(),
"cache_enabled": torch.is_autocast_cache_enabled()}
cpu_autocast_kwargs = {"enabled": torch.is_autocast_cpu_enabled(),
"dtype": torch.get_autocast_cpu_dtype(),
"cache_enabled": torch.is_autocast_cache_enabled()}
return gpu_autocast_kwargs, cpu_autocast_kwargs
def __enter__(self):
if self.device == 'cpu':
self.prev = torch.is_autocast_cpu_enabled()
self.prev_fastdtype = torch.get_autocast_cpu_dtype()
torch.set_autocast_cpu_enabled(self._enabled)
torch.set_autocast_cpu_dtype(self.fast_dtype)
torch.autocast_increment_nesting()
else:
self.prev = torch.is_autocast_enabled()
self.prev_fastdtype = torch.get_autocast_gpu_dtype()
torch.set_autocast_gpu_dtype(self.fast_dtype)
torch.set_autocast_enabled(self._enabled)
torch.autocast_increment_nesting()
def _validate_inputs(self, a, b):
if a.device != b.device:
raise ValueError(
f"Inputs must be on the same device; got {a.device} for tensor A "
f"and {b.device} for tensor B")
if not a.is_cuda:
raise ValueError("Only GPU devices are supported for now")
# When autocast is enabled, torch.matmul autocasts to float16, so we do the same here
if torch.is_autocast_enabled():
a, b = a.half(), b.half()
elif a.dtype != b.dtype:
raise ValueError(
f"Inputs must be the same dtype; got {a.dtype} for A and {b.dtype} for B"
)
mode, trans_a, trans_b = self.mode, self.trans_a, self.trans_b
if mode != 'sdd':
# One input is sparse
dense, dense_name, sparse, sparse_name = (
a, 'A', b, 'B') if mode == 'dds' else (b, 'B', a, 'A')
dense_inner = dense.shape[self.dense_inner_dim]
if dense_inner != self.dense_inner_size:
raise ValueError(
f"Expected tensor {dense_name} to have size {self.dense_inner_size} at dim "
f"{self.dense_inner_dim % dense.ndim}, got {dense_inner}.")
if sparse.shape[-len(self.sparse_shape):] != self.sparse_shape:
raise ValueError(
f"Expected tensor with trailing dimensions of shape {self.sparse_shape} for argument "
f"{sparse_name}, got {sparse.shape}")
def add_extra_dims(x):
# Add extra leading singleton dimensions if needed
dims_needed = 4 - x.ndim
if dims_needed > 0:
singletons = [1] * dims_needed
x = x.view(*singletons, *x.shape)
elif dims_needed < 0:
raise ValueError(
"Tensors with more than 4 dimensions are not currently supported"
)
return x
# Pad shapes with leading singleton dimensions
a = add_extra_dims(a)
b = add_extra_dims(b)
return a, b
def causal_linear_attention(q, k, v, eps = 1e-6):
from fast_transformers.causal_product import CausalDotProduct
autocast_enabled = torch.is_autocast_enabled()
is_half = isinstance(q, torch.cuda.HalfTensor)
assert not is_half or APEX_AVAILABLE, 'half tensors can only be used if nvidia apex is available'
cuda_context = null_context if not autocast_enabled else partial(autocast, enabled = False)
causal_dot_product_fn = amp.float_function(CausalDotProduct.apply) if is_half else CausalDotProduct.apply
k_cumsum = k.cumsum(dim=-2) + eps
D_inv = 1. / torch.einsum('...nd,...nd->...n', q, k_cumsum.type_as(q))
with cuda_context():
if autocast_enabled:
q, k, v = map(lambda t: t.float(), (q, k, v))
out = causal_dot_product_fn(q, k, v)
out = torch.einsum('...nd,...n->...nd', out, D_inv)
return out
def forward(ctx, run_function, preserve_rng_state, *args):
check_backward_validity(args)
ctx.run_function = run_function
ctx.preserve_rng_state = preserve_rng_state
ctx.had_autocast_in_fwd = torch.is_autocast_enabled()
if preserve_rng_state:
ctx.fwd_cpu_state = torch.get_rng_state()
# Don't eagerly initialize the cuda context by accident.
# (If the user intends that the context is initialized later, within their
# run_function, we SHOULD actually stash the cuda state here. Unfortunately,
# we have no way to anticipate this will happen before we run the function.)
ctx.had_cuda_in_fwd = False
if torch.cuda._initialized:
ctx.had_cuda_in_fwd = True
ctx.fwd_gpu_devices, ctx.fwd_gpu_states = get_device_states(
*args)
ctx.save_for_backward(*args)
with torch.no_grad():
outputs = run_function(*args)
return outputs
def attention(query, key, value, mask=None, key_padding_mask=None, dropout=None,fixed=False,att_bias=None):
#print(query.size()) #batch,heads,len,feats
"Compute 'Scaled Dot Product Attention'"
d_k = query.size(-1)
#bsz, num_heads, src_len,d_k = key.size()
scores = torch.matmul(query, key.transpose(-2, -1)) \
/ math.sqrt(d_k)
###
#scores.fill_(0.1)
###
#scores.size() = (batch heads queries keys)
if att_bias is not None:
scores+=att_bias
if mask is not None:
if torch.is_autocast_enabled():
scores = scores.masked_fill(mask == 0, -1e4)
else:
scores = scores.masked_fill(mask == 0, -1e9)
if key_padding_mask is not None:
#import pdb;pdb.set_trace()
#tgt_len = query.size(2)
#scores = scores.view(bsz, num_heads, tgt_len, src_len
scores = scores.masked_fill(
key_padding_mask.unsqueeze(1).unsqueeze(2),
float("-inf"),
)
#scores = scores.view(bsz * num_heads, tgt_len, src_len)
p_attn = F.softmax(scores, dim = -1)
if mask is not None and fixed:
p_attn = p_attn.masked_fill(mask == 0, 0) #this is needed in case a node has no neigbors (softmax gives even attention to everything
#will create a zero vector in those cases, instead of the average of all nodes
if dropout is not None:
p_attn = dropout(p_attn)
return torch.matmul(p_attn, value), p_attn
def _get_current_dtype(dtype: Optional[torch.dtype] = None) -> torch.dtype:
if not torch.is_autocast_enabled():
return torch.float or dtype
else:
return torch.get_autocast_gpu_dtype()
def _run_autocast_outofplace(self,
op,
args,
run_as_type,
out_type=None,
module=torch,
add_kwargs=None):
# helper to cast args
def cast(val, to_type):
if isinstance(val, torch.Tensor):
return val.to(to_type) if val.is_floating_point() else val
elif isinstance(val, collections.abc.Iterable):
return type(val)(cast(v, to_type) for v in val)
else:
return val
if add_kwargs is None:
add_kwargs = {}
self.assertFalse(torch.is_autocast_enabled())
with autocast():
self.assertTrue(torch.is_autocast_enabled())
out_type = out_type if out_type is not None else run_as_type
output = output_method = None
# Try module.* variant, if requested:
if module is not None and hasattr(module, op):
output = getattr(module, op)(*args, **add_kwargs)
if isinstance(output, torch.Tensor):
self.assertTrue(
out_type == output.dtype,
"autocast for torch.{} produced {}, should produce {}".
format(op, output.dtype, out_type))
# Try Tensor.* variant:
if hasattr(torch.Tensor, op):
output_method = getattr(args[0], op)(*args[1:], **add_kwargs)
if isinstance(output_method, torch.Tensor):
self.assertTrue(
out_type == output_method.dtype,
"autocast for torch.{} produced {}, should produce torch.{}"
.format(op, output_method.dtype, out_type))
self.assertTrue((output is not None) or (
output_method is not None
), "{} not found as an attribute on either Tensor or the requested module {}"
.format(op, module))
# Accounts for ops that return Tensors, iterables, and other non-Tensors.
# For example, lstm_cell returns a tuple and equal returns bool.
def compare(first, second):
if isinstance(first, torch.Tensor):
return torch.equal(first, second)
elif isinstance(first, collections.abc.Iterable):
return all(compare(f, s) for f, s in zip(first, second))
else:
return first == second
# If both torch.* and Tensor.* variants were found, check outputs are identical
if (output is not None) and (output_method is not None):
self.assertTrue(type(output) == type(output_method))
comparison = compare(output, output_method)
self.assertTrue(
comparison,
"torch.{0} result did not match Tensor.{0} result".format(
op))
# Compare numerics to Python-side "autocasting" that (we expect) does the same thing
# as the C++-side autocasting, and should be bitwise accurate.
output_to_compare = output if output is not None else output_method
with autocast(enabled=False):
self.assertFalse(torch.is_autocast_enabled())
if module is not None and hasattr(module, op):
control = getattr(module, op)(*cast(args, run_as_type),
**add_kwargs)
else:
control = getattr(args[0].to(run_as_type),
op)(*cast(args[1:], run_as_type),
**add_kwargs)
self.assertTrue(type(output_to_compare) == type(control))
comparison = compare(output_to_compare, control)
self.assertTrue(
comparison,
"torch.{} result did not match control".format(op))
self.assertTrue(torch.is_autocast_enabled())
self.assertFalse(torch.is_autocast_enabled())
def __enter__(self):
self.prev = torch.is_autocast_enabled()
torch.set_autocast_enabled(self._enabled)
torch.autocast_increment_nesting()
def _checkpoint_without_reentrant(function, preserve_rng_state=True, *args):
"""Checkpointining without re-entrant autograd
Args:
function: describes what to run in the forward pass of the model or
part of the model. It should also know how to handle the inputs
passed as the tuple. For example, in LSTM, if user passes
``(activation, hidden)``, :attr:`function` should correctly use the
first input as ``activation`` and the second input as ``hidden``
preserve_rng_state(bool, optional, default=True): Omit stashing and restoring
the RNG state during each checkpoint.
*args: Arguments to pass in to the given ``function``.
"""
had_autocast_in_fwd = torch.is_autocast_enabled()
if preserve_rng_state:
fwd_cpu_state = torch.get_rng_state()
# Don't eagerly initialize the cuda context by accident.
# (If the user intends that the context is initialized later, within their
# run_function, we SHOULD actually stash the cuda state here. Unfortunately,
# we have no way to anticipate this will happen before we run the function.
# If they do so, we raise an error.)
had_cuda_in_fwd = False
if torch.cuda._initialized:
had_cuda_in_fwd = True
fwd_gpu_devices, fwd_gpu_states = get_device_states(*args)
storage: List[Union[torch.Tensor, None]] = []
counter = 0
def pack(x):
nonlocal counter
counter += 1
# TODO(varal7): Instead of returning indices, we can return things metadata (such as
# size, device, ...) to catch certain cases of undeterministic behavior of the forward
return counter - 1
def unpack(x):
if len(storage) == 0:
def inner_pack(inner):
storage.append(inner)
return None
def inner_unpack(packed):
raise RuntimeError(
"You are calling backwards on a tensor that is never exposed. Please open an issue."
)
# Stash the surrounding rng state, and mimic the state that was
# present at this time during forward. Restore the surrounding state
# when we're done.
rng_devices = []
if preserve_rng_state and had_cuda_in_fwd:
rng_devices = fwd_gpu_devices
with torch.random.fork_rng(devices=rng_devices,
enabled=preserve_rng_state):
if preserve_rng_state:
torch.set_rng_state(fwd_cpu_state)
if had_cuda_in_fwd:
set_device_states(fwd_gpu_devices, fwd_gpu_states)
with torch.enable_grad(), torch.cuda.amp.autocast(
had_autocast_in_fwd):
with torch.autograd.graph.saved_tensors_hooks(
inner_pack, inner_unpack):
_unused = function(*args)
return storage[x]
with torch.autograd.graph.saved_tensors_hooks(pack, unpack):
output = function(*args)
if torch.cuda._initialized and not had_cuda_in_fwd:
# Cuda was not initialized before running the forward, so we didn't
# stash the CUDA state.
raise RuntimeError(
"PyTorch's CUDA state was initialized in the forward pass "
"of a Checkpoint, which is not allowed. Please open an issue "
"if you need this feature.")
return output
def _cast_if_autocast_enabled(*args):
if not torch.is_autocast_enabled():
return args
else:
return torch.cuda.amp.autocast_mode._cast(args, torch.get_autocast_gpu_dtype())
def is_autocast_enabled() -> bool:
"""Similar to torch.is_autocast_enabled, but compatible with torch 1.5.1"""
if hasattr(torch, "is_autocast_enabled"):
return torch.is_autocast_enabled()
return False
def forward(self, x):
"""Forward pass"""
return self.mod.forward(x.to(torch.float).to(
x.dtype)) if torch.is_autocast_enabled() else self.mod.forward(x)
def _assert_autocast_enabled(self):
if self.trainer.precision_plugin.device == "cpu":
assert torch.is_autocast_cpu_enabled()
else:
assert torch.is_autocast_enabled()
def parallel_apply(modules, inputs, kwargs_tup=None, devices=None):
r"""Applies each `module` in :attr:`modules` in parallel on arguments
contained in :attr:`inputs` (positional) and :attr:`kwargs_tup` (keyword)
on each of :attr:`devices`.
Args:
modules (Module): modules to be parallelized
inputs (tensor): inputs to the modules
devices (list of int or torch.device): CUDA devices
:attr:`modules`, :attr:`inputs`, :attr:`kwargs_tup` (if given), and
:attr:`devices` (if given) should all have same length. Moreover, each
element of :attr:`inputs` can either be a single object as the only argument
to a module, or a collection of positional arguments.
"""
assert len(modules) == len(inputs)
if kwargs_tup is not None:
assert len(modules) == len(kwargs_tup)
else:
kwargs_tup = ({}, ) * len(modules)
if devices is not None:
assert len(modules) == len(devices)
else:
devices = [None] * len(modules)
devices = list(map(lambda x: _get_device_index(x, True), devices))
lock = threading.Lock()
results = {}
grad_enabled, autocast_enabled = torch.is_grad_enabled(
), torch.is_autocast_enabled()
def _worker(i, module, input, kwargs, device=None):
torch.set_grad_enabled(grad_enabled)
if device is None:
device = get_a_var(input).get_device()
try:
with torch.cuda.device(device), autocast(enabled=autocast_enabled):
# this also avoids accidental slicing of `input` if it is a Tensor
if not isinstance(input, (list, tuple)):
input = (input, )
output = module(*input, **kwargs)
with lock:
results[i] = output
except Exception:
with lock:
results[i] = ExceptionWrapper(
where="in replica {} on device {}".format(i, device))
if len(modules) > 1:
threads = [
threading.Thread(target=_worker,
args=(i, module, input, kwargs, device))
for i, (
module, input, kwargs,
device) in enumerate(zip(modules, inputs, kwargs_tup, devices))
]
for thread in threads:
thread.start()
for thread in threads:
thread.join()
else:
_worker(0, modules[0], inputs[0], kwargs_tup[0], devices[0])
outputs = []
for i in range(len(inputs)):
output = results[i]
if isinstance(output, ExceptionWrapper):
output.reraise()
outputs.append(output)
return outputs
def _step(self, batch, batch_idx):
assert torch.is_autocast_enabled()
output = self(batch)
assert output.dtype == torch.float16
loss = self.loss(batch, output)
return loss
def check_autocast(forward_input):
assert precision != 16 or torch.is_autocast_enabled()
return forward_input
def predict(self, batch, batch_idx, dataloader_idx=None):
assert torch.is_autocast_enabled()
output = self(batch)
assert output.dtype == torch.float16
return output
def to_dtype(x, tensor=None, dtype=None):
if not torch.is_autocast_enabled():
dt = dtype if dtype is not None else tensor.dtype
if x.dtype != dt:
x = x.type(dt)
return x