Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Dygraph]optimizer sharding paramters #39581

Merged
merged 1 commit into from
Feb 17, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -65,9 +65,9 @@ def __init__(self,
params,
optim,
group=None,
broadcast_fp16=False,
offload=False,
device="gpu",
pertrain_sync_models=True,
**kw):

super().__init__(optim._learning_rate, params, kw)
Expand Down Expand Up @@ -98,8 +98,12 @@ def __init__(self,

self.world_size = self.group.nranks
self.rank = self.group.rank
self._global_root_rank = 0

# Synchronous all ranks models
if pertrain_sync_models:
self._sync_params_and_buffers()

self.broadcast_fp16 = broadcast_fp16
self.param_storages = {} # {dtype: {rank: InternalStorage}}

if isinstance(self._optim._grad_clip, ClipGradByGlobalNorm):
Expand Down Expand Up @@ -132,6 +136,22 @@ def __init__(self,
# Update optimizer parameters and adjust parameter storage and use according to rank.
self._update_opt_status()

@paddle.no_grad()
def _sync_params_and_buffers(self):
"""
Sync all model states for all ranks
"""

for p in self._local_params:
dist.broadcast(
p,
src=self._global_root_rank,
group=self.group,
use_calc_stream=True)

# Multi stream operation will be supported later
dist.wait(tensor=p, group=self.group, use_calc_stream=True)

def _generate_master_params(self, trainable_params):
if self.offload:
for param in trainable_params:
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -61,12 +61,10 @@ def __init__(
sharding_optimizer,
group=None,
sync_buffers=False,
pertrain_sync_models=True,
buffer_max_size=2**23, #8MB
auto_refresh_trainable=True,
device="gpu",
use_grad_storage=True,
accumulate_grads=False):
use_grad_storage=True):
super().__init__()

# training options
Expand All @@ -81,9 +79,6 @@ def __init__(
self._sync_buffers = sync_buffers
self._auto_refresh_trainable = auto_refresh_trainable

# Gradient accumulation, Gradient flip
self._accumulate_grads = accumulate_grads

# Communication related attributes
self._group = dist.new_group(_get_global_group()
.ranks) if group is None else group
Expand Down Expand Up @@ -128,16 +123,11 @@ def __init__(
# Set backward pass hooks
self._bw_hooks = []

# Synchronous all ranks models
if pertrain_sync_models:
self._sync_params_and_buffers()

# Set tasks flow
self._tasks_flow = deque()

# Define optimizer step and clear_grad
if self._accumulate_grads:
self._redefine_opt_step()
self._redefine_opt_step()
self._redefine_opt_clear()

def forward(self, *inputs, **kwargs):
Expand Down Expand Up @@ -313,9 +303,6 @@ def reduce(*_):

# Change reduce information
self._grad_reduced[index] = False
if not self._accumulate_grads:
param.grad.scale_(scale=self._world_size_scaling)
param._reset_grad_inplace_version(True)

# Clear the gradient that does not belong to the current rank through the callback function
def cleanup():
Expand Down Expand Up @@ -362,11 +349,6 @@ def reduce(*_):
if grad_storage.all_checked_in:
assert grad_storage.buffer is not None

# Normalize all ranks grad_storage
if not self._accumulate_grads:
grad_storage.buffer.scale_(
scale=self._world_size_scaling)

# Clearing up the grad_storage buffer
def cleanup():
if dst_rank != self._rank:
Expand Down Expand Up @@ -432,22 +414,6 @@ def _setup_backward_hooks(self):
self._bw_hooks.append(
param._register_backward_hook(reduce_function))

@paddle.no_grad()
def _sync_params_and_buffers(self):
"""
Sync all model states for all ranks
"""

for t in self._layer.parameters():
dist.broadcast(
t,
src=self._global_root_rank,
group=self._group,
use_calc_stream=True)

# Multi stream operation will be supported later
dist.wait(tensor=t, group=self._group, use_calc_stream=True)

def _setup_use_grad_storage(self):
"""
Integrate the parameters gradient into a continuous memory according to rank, and support the update of training parameters.
Expand Down Expand Up @@ -555,8 +521,6 @@ def _rank_buffer_size(self, buffer_max_size, model_size):
return rank_buffer_size

def _redefine_opt_step(self):
if not self._accumulate_grads:
return
grad_func = self._grad_scale
for opt in self._sharding_optimizers:
opt_step = opt.step
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -72,7 +72,6 @@ def __init__(self,
device="gpu",
segment_size=2**15,
pertrain_sync_models=True,
accumulate_grads=False,
offload=False,
sync_comm=False):
super().__init__()
Expand All @@ -82,7 +81,6 @@ def __init__(self,
self._layer = layer
self._default_device = device
self.__sync_buffers = sync_buffers
self._accumulate_grads = accumulate_grads
self._offload = offload
self._sync_comm = sync_comm
# segmentation size
Expand Down Expand Up @@ -190,6 +188,7 @@ def _clear_gradients(self):
param.fw_storage.clear_gradient(False)
param.fw_storage._gradient_set_empty(False)
param.bw_storage._clear()
param.bw_storage = None
# 2.Handle unslice param
if not self._offload:
for grad_storage in self._grad_storages.values():
Expand Down Expand Up @@ -446,13 +445,12 @@ def _update_params(self):
param,
"fw_storage"), "Find {} don't have fw_storage attribute".format(
param.name)

if self._accumulate_grads:
if self._offload:
with device_guard(device="cpu"):
param.bw_storage.scale_(scale=self._world_size_scaling)
else:
# Gradient average
if self._offload:
with device_guard(device="cpu"):
param.bw_storage.scale_(scale=self._world_size_scaling)
else:
param.bw_storage.scale_(scale=self._world_size_scaling)
param.fw_storage = _VarBaseWrapper(param)
assert param.fw_storage.grad is None
param.fw_storage._copy_gradient_from(param.bw_storage)
Expand Down Expand Up @@ -526,17 +524,14 @@ def _get_allreduce_fn(self, param):
def reduce(*_):
if param.name in self._task_flow.full_grad.keys():
full_grad = self._task_flow.full_grad[param.name]
if not self._accumulate_grads:
full_grad.scale_(scale=self._world_size_scaling)
# Only support sync allreduce current rank's layer now
dist.all_reduce(
tensor=full_grad, group=self._group, use_calc_stream=True)
dist.wait(
tensor=full_grad, group=self._group, use_calc_stream=True)

start, end = self._param2buffer[param.name][self._rank]
if not self._accumulate_grads or param.bw_storage is None or not param.bw_storage.value(
).get_tensor()._is_initialized():
if param.bw_storage is None:
param.bw_storage = core.VarBase(
full_grad._slice(start, end)).detach().clone()
if self._offload:
Expand Down
15 changes: 8 additions & 7 deletions python/paddle/fluid/tests/unittests/dygraph_sharding_stage2.py
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@
from paddle.distributed.fleet.meta_optimizers.dygraph_optimizer.sharding_optimizer_stage2 import ShardingOptimizerStage2
from paddle.distributed.fleet.meta_parallel.sharding.sharding_stage2 import ShardingStage2

seed = 2021
seed = 2022
epoch = 2
linear_size = 1000

Expand Down Expand Up @@ -105,11 +105,7 @@ def train_mlp(model,
params=model.parameters(), optim=optimizer, group=group)

model = ShardingStage2(
model,
optimizer,
group=group,
buffer_max_size=2**21,
accumulate_grads=batch_size == 20)
model, optimizer, group=group, buffer_max_size=2**21)
else:
optimizer = fleet.distributed_optimizer(optimizer)
model = fleet.distributed_model(model)
Expand Down Expand Up @@ -140,6 +136,8 @@ def train_mlp(model,
loss = paddle.nn.functional.cross_entropy(input=out, label=label)

avg_loss = paddle.mean(x=loss.cast(dtype=paddle.float32))
if batch_size == 20:
avg_loss = avg_loss / 5
avg_loss.backward()

if not accumulate_grad:
Expand All @@ -166,6 +164,7 @@ def test_dp_stage2():
mlp4.set_state_dict(state_dict)
mlp5.set_state_dict(state_dict)

# DP VS stage2
dp_params = train_mlp(
mlp1, sharding_stage="dp", use_pure_fp16=False, opt_group=False)
stage2_params = train_mlp(
Expand All @@ -174,7 +173,8 @@ def test_dp_stage2():
np.testing.assert_allclose(
dp_params[i].numpy(), stage2_params[i].numpy(), rtol=1e-6)

stage2_params = train_mlp(mlp3, sharding_stage=2)
# stage2 accumulate grad
stage2_params = train_mlp(mlp3, sharding_stage=2, accumulate_grad=True)
stage2_accumulate_grad = train_mlp(
mlp4, sharding_stage=2, batch_size=20, accumulate_grad=True)
for i in range(len(stage2_params)):
Expand All @@ -184,6 +184,7 @@ def test_dp_stage2():
rtol=1e-5,
atol=1e-5)

# stage2 param list VS param group
stage2_params = train_mlp(
mlp2, sharding_stage=2, use_pure_fp16=False, opt_group=True)
for i in range(len(dp_params)):
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -43,13 +43,12 @@ def train_mlp(model, offload=False):
optimizer = optimizer_setting(model=model, use_pure_fp16=True)

model = paddle.amp.decorate(models=model, level='O2', save_dtype='float32')
scaler = paddle.amp.GradScaler(init_loss_scaling=32768)
scaler = paddle.amp.GradScaler(init_loss_scaling=1024)
scaler = ShardingScaler(scaler)

optimizer = ShardingOptimizerStage2(
params=model.parameters(), optim=optimizer, offload=offload)
model = ShardingStage2(
model, optimizer, buffer_max_size=2**21, accumulate_grads=False)
model = ShardingStage2(model, optimizer, buffer_max_size=2**21)

train_reader = paddle.batch(
reader_decorator(linear_size), batch_size=batch_size, drop_last=True)
Expand Down
14 changes: 3 additions & 11 deletions python/paddle/fluid/tests/unittests/dygraph_sharding_stage3.py
Original file line number Diff line number Diff line change
Expand Up @@ -101,18 +101,10 @@ def train_mlp(model,
optimizer = ShardingOptimizerStage2(
params=model.parameters(), optim=optimizer, group=group)
model = ShardingStage2(
model,
optimizer,
group=group,
buffer_max_size=2**21,
accumulate_grads=batch_size == 20)
model, optimizer, group=group, buffer_max_size=2**21)
elif sharding_stage == 3:
model = ShardingStage3(
model,
optimizer=optimizer,
group=group,
accumulate_grads=batch_size == 20,
sync_comm=recompute)
model, optimizer=optimizer, group=group, sync_comm=recompute)

# check optimizer.minimize() error
if test_minimize:
Expand Down Expand Up @@ -231,7 +223,7 @@ def test_stage2_stage3():
stage2_params[i].numpy(),
stage3_params[i].numpy(),
rtol=1e-4,
atol=1e-4)
atol=1e-3)

# fp16 recompute
stage3_params = train_mlp(
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -91,11 +91,7 @@ def train_mlp(model,
scaler = ShardingScaler(scaler)

model = ShardingStage3(
model,
optimizer=optimizer,
group=group,
offload=offload,
accumulate_grads=accumulate_grad)
model, optimizer=optimizer, group=group, offload=offload)

train_reader = paddle.batch(
reader_decorator(), batch_size=batch_size, drop_last=True)
Expand Down