Traning Big Models¶
When training large models, significant resources are required. A single GPU memory is often insufficient to meet the training needs. As a result, techniques for training large models have been developed, and one typical approach is DeepSpeed ZeRO. DeepSpeed ZeRO supports optimizer, gradient, and parameter sharding.
To provide more flexibility in supporting large model training techniques, starting from MMEngine v0.8.0, we have introduced a new runner called FlexibleRunner and multiple abstract Strategies.
Warning
The new FlexibleRunner and Strategy are still in the experimental stage, and their interfaces may change in future versions.
The following example code is excerpted from examples/distributed_training_with_flexible_runner.py.
DeepSpeed¶
DeepSpeed is an open-source distributed framework based on PyTorch, developed by Microsoft. It supports training strategies such as ZeRO
, 3D-Parallelism
, DeepSpeed-MoE
, and ZeRO-Infinity
.
Starting from MMEngine v0.8.0, MMEngine supports training models using DeepSpeed.
To use DeepSpeed, you need to install it first by running the following command:
pip install deepspeed
After installing DeepSpeed, you need to configure the strategy
and optim_wrapper
parameters of FlexibleRunner as follows:
strategy: Set
type='DeepSpeedStrategy'
and configure other parameters. See DeepSpeedStrategy for more details.optim_wrapper: Set
type='DeepSpeedOptimWrapper'
and configure other parameters. See DeepSpeedOptimWrapper for more details.
Here is an example configuration related to DeepSpeed:
from mmengine.runner._flexible_runner import FlexibleRunner
# set `type='DeepSpeedStrategy'` and configure other parameters
strategy = dict(
type='DeepSpeedStrategy',
fp16=dict(
enabled=True,
fp16_master_weights_and_grads=False,
loss_scale=0,
loss_scale_window=500,
hysteresis=2,
min_loss_scale=1,
initial_scale_power=15,
),
inputs_to_half=[0],
zero_optimization=dict(
stage=3,
allgather_partitions=True,
reduce_scatter=True,
allgather_bucket_size=50000000,
reduce_bucket_size=50000000,
overlap_comm=True,
contiguous_gradients=True,
cpu_offload=False),
)
# set `type='DeepSpeedOptimWrapper'` and configure other parameters
optim_wrapper = dict(
type='DeepSpeedOptimWrapper',
optimizer=dict(type='AdamW', lr=1e-3))
# construct FlexibleRunner
runner = FlexibleRunner(
model=MMResNet50(),
work_dir='./work_dirs',
strategy=strategy,
train_dataloader=train_dataloader,
optim_wrapper=optim_wrapper,
param_scheduler=dict(type='LinearLR'),
train_cfg=dict(by_epoch=True, max_epochs=10, val_interval=1),
val_dataloader=val_dataloader,
val_cfg=dict(),
val_evaluator=dict(type=Accuracy))
# start training
runner.train()
Using two GPUs to launch distributed training:
torchrun --nproc-per-node 2 examples/distributed_training_with_flexible_runner.py --use-deepspeed
traning log
07/03 13:04:17 - mmengine - INFO - Epoch(train) [1][ 10/196] lr: 3.3333e-04 eta: 0:13:14 time: 0.4073 data_time: 0.0335 memory: 970 loss: 6.1887
07/03 13:04:19 - mmengine - INFO - Epoch(train) [1][ 20/196] lr: 3.3333e-04 eta: 0:09:39 time: 0.1904 data_time: 0.0327 memory: 970 loss: 2.5746
07/03 13:04:21 - mmengine - INFO - Epoch(train) [1][ 30/196] lr: 3.3333e-04 eta: 0:08:32 time: 0.1993 data_time: 0.0342 memory: 970 loss: 2.4180
07/03 13:04:23 - mmengine - INFO - Epoch(train) [1][ 40/196] lr: 3.3333e-04 eta: 0:08:01 time: 0.2052 data_time: 0.0368 memory: 970 loss: 2.3682
07/03 13:04:25 - mmengine - INFO - Epoch(train) [1][ 50/196] lr: 3.3333e-04 eta: 0:07:39 time: 0.2013 data_time: 0.0356 memory: 970 loss: 2.3025
07/03 13:04:27 - mmengine - INFO - Epoch(train) [1][ 60/196] lr: 3.3333e-04 eta: 0:07:25 time: 0.2025 data_time: 0.0353 memory: 970 loss: 2.2078
07/03 13:04:29 - mmengine - INFO - Epoch(train) [1][ 70/196] lr: 3.3333e-04 eta: 0:07:13 time: 0.1999 data_time: 0.0352 memory: 970 loss: 2.2045
07/03 13:04:31 - mmengine - INFO - Epoch(train) [1][ 80/196] lr: 3.3333e-04 eta: 0:07:04 time: 0.2013 data_time: 0.0350 memory: 970 loss: 2.1709
07/03 13:04:33 - mmengine - INFO - Epoch(train) [1][ 90/196] lr: 3.3333e-04 eta: 0:06:56 time: 0.1975 data_time: 0.0341 memory: 970 loss: 2.2070
07/03 13:04:35 - mmengine - INFO - Epoch(train) [1][100/196] lr: 3.3333e-04 eta: 0:06:49 time: 0.1993 data_time: 0.0347 memory: 970 loss: 2.0891
07/03 13:04:37 - mmengine - INFO - Epoch(train) [1][110/196] lr: 3.3333e-04 eta: 0:06:44 time: 0.1995 data_time: 0.0357 memory: 970 loss: 2.0700
07/03 13:04:39 - mmengine - INFO - Epoch(train) [1][120/196] lr: 3.3333e-04 eta: 0:06:38 time: 0.1966 data_time: 0.0342 memory: 970 loss: 1.9983
07/03 13:04:41 - mmengine - INFO - Epoch(train) [1][130/196] lr: 3.3333e-04 eta: 0:06:37 time: 0.2216 data_time: 0.0341 memory: 970 loss: 1.9409
07/03 13:04:43 - mmengine - INFO - Epoch(train) [1][140/196] lr: 3.3333e-04 eta: 0:06:32 time: 0.1944 data_time: 0.0336 memory: 970 loss: 1.9800
07/03 13:04:45 - mmengine - INFO - Epoch(train) [1][150/196] lr: 3.3333e-04 eta: 0:06:27 time: 0.1946 data_time: 0.0338 memory: 970 loss: 1.9356
07/03 13:04:47 - mmengine - INFO - Epoch(train) [1][160/196] lr: 3.3333e-04 eta: 0:06:22 time: 0.1937 data_time: 0.0333 memory: 970 loss: 1.8145
07/03 13:04:49 - mmengine - INFO - Epoch(train) [1][170/196] lr: 3.3333e-04 eta: 0:06:18 time: 0.1941 data_time: 0.0335 memory: 970 loss: 1.8525
07/03 13:04:51 - mmengine - INFO - Epoch(train) [1][180/196] lr: 3.3333e-04 eta: 0:06:17 time: 0.2204 data_time: 0.0341 memory: 970 loss: 1.7637
07/03 13:04:53 - mmengine - INFO - Epoch(train) [1][190/196] lr: 3.3333e-04 eta: 0:06:14 time: 0.1998 data_time: 0.0345 memory: 970 loss: 1.7523
FullyShardedDataParallel (FSDP)¶
PyTorch has supported training with FullyShardedDataParallel (FSDP) since version v1.11. However, due to its evolving interface, we only support PyTorch versions 2.0.0 and above.
To use FSDP, you need to configure the 'strategy'
parameter of FlexibleRunner by specifying type='FSDPStrategy'
and configuring the parameters. For detailed information about it, you can refer to FSDPStrategy.
Here is an example configuration related to FSDP:
from torch.distributed.fsdp.wrap import size_based_auto_wrap_policy
size_based_auto_wrap_policy = partial(
size_based_auto_wrap_policy, min_num_params=1e7)
# set `type='FSDPStrategy'` and configure other parameters
strategy = dict(
type='FSDPStrategy',
model_wrapper=dict(auto_wrap_policy=size_based_auto_wrap_policy))
# set `type='AmpOptimWrapper'` and configure other parameters
optim_wrapper = dict(
type='AmpOptimWrapper', optimizer=dict(type='AdamW', lr=1e-3))
# construct FlexibleRunner
runner = FlexibleRunner(
model=MMResNet50(),
work_dir='./work_dirs',
strategy=strategy,
train_dataloader=train_dataloader,
optim_wrapper=optim_wrapper,
param_scheduler=dict(type='LinearLR'),
train_cfg=dict(by_epoch=True, max_epochs=10, val_interval=1),
val_dataloader=val_dataloader,
val_cfg=dict(),
val_evaluator=dict(type=Accuracy))
# start training
runner.train()
Using two GPUs to launch distributed training:
torchrun --nproc-per-node 2 examples/distributed_training_with_flexible_runner.py --use-fsdp
traning log
07/03 13:05:37 - mmengine - INFO - Epoch(train) [1][ 10/196] lr: 3.3333e-04 eta: 0:08:28 time: 0.2606 data_time: 0.0330 memory: 954 loss: 6.1265
07/03 13:05:38 - mmengine - INFO - Epoch(train) [1][ 20/196] lr: 3.3333e-04 eta: 0:05:18 time: 0.0673 data_time: 0.0325 memory: 954 loss: 2.5584
07/03 13:05:39 - mmengine - INFO - Epoch(train) [1][ 30/196] lr: 3.3333e-04 eta: 0:04:13 time: 0.0666 data_time: 0.0320 memory: 954 loss: 2.4816
07/03 13:05:39 - mmengine - INFO - Epoch(train) [1][ 40/196] lr: 3.3333e-04 eta: 0:03:41 time: 0.0666 data_time: 0.0321 memory: 954 loss: 2.3695
07/03 13:05:40 - mmengine - INFO - Epoch(train) [1][ 50/196] lr: 3.3333e-04 eta: 0:03:21 time: 0.0671 data_time: 0.0324 memory: 954 loss: 2.3208
07/03 13:05:41 - mmengine - INFO - Epoch(train) [1][ 60/196] lr: 3.3333e-04 eta: 0:03:08 time: 0.0667 data_time: 0.0320 memory: 954 loss: 2.2431
07/03 13:05:41 - mmengine - INFO - Epoch(train) [1][ 70/196] lr: 3.3333e-04 eta: 0:02:58 time: 0.0667 data_time: 0.0320 memory: 954 loss: 2.1873
07/03 13:05:42 - mmengine - INFO - Epoch(train) [1][ 80/196] lr: 3.3333e-04 eta: 0:02:51 time: 0.0669 data_time: 0.0320 memory: 954 loss: 2.2006
07/03 13:05:43 - mmengine - INFO - Epoch(train) [1][ 90/196] lr: 3.3333e-04 eta: 0:02:45 time: 0.0671 data_time: 0.0324 memory: 954 loss: 2.1547
07/03 13:05:43 - mmengine - INFO - Epoch(train) [1][100/196] lr: 3.3333e-04 eta: 0:02:40 time: 0.0667 data_time: 0.0321 memory: 954 loss: 2.1361
07/03 13:05:44 - mmengine - INFO - Epoch(train) [1][110/196] lr: 3.3333e-04 eta: 0:02:36 time: 0.0668 data_time: 0.0320 memory: 954 loss: 2.0405
07/03 13:05:45 - mmengine - INFO - Epoch(train) [1][120/196] lr: 3.3333e-04 eta: 0:02:32 time: 0.0669 data_time: 0.0320 memory: 954 loss: 2.0228
07/03 13:05:45 - mmengine - INFO - Epoch(train) [1][130/196] lr: 3.3333e-04 eta: 0:02:29 time: 0.0670 data_time: 0.0324 memory: 954 loss: 2.0375
07/03 13:05:46 - mmengine - INFO - Epoch(train) [1][140/196] lr: 3.3333e-04 eta: 0:02:26 time: 0.0664 data_time: 0.0320 memory: 954 loss: 1.9926
07/03 13:05:47 - mmengine - INFO - Epoch(train) [1][150/196] lr: 3.3333e-04 eta: 0:02:24 time: 0.0668 data_time: 0.0320 memory: 954 loss: 1.9820
07/03 13:05:47 - mmengine - INFO - Epoch(train) [1][160/196] lr: 3.3333e-04 eta: 0:02:22 time: 0.0674 data_time: 0.0325 memory: 954 loss: 1.9728
07/03 13:05:48 - mmengine - INFO - Epoch(train) [1][170/196] lr: 3.3333e-04 eta: 0:02:20 time: 0.0666 data_time: 0.0320 memory: 954 loss: 1.9359
07/03 13:05:49 - mmengine - INFO - Epoch(train) [1][180/196] lr: 3.3333e-04 eta: 0:02:18 time: 0.0667 data_time: 0.0321 memory: 954 loss: 1.9488
07/03 13:05:49 - mmengine - INFO - Epoch(train) [1][190/196] lr: 3.3333e-04 eta: 0:02:16 time: 0.0671 data_time: 0.0323 memory: 954 loss: 1.9023\