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.

Sign up for GitHub

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

feat: add support for tensor parallel using Pytorch #34194

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

feat: add support for tensor parallel using Pytorch #34194

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
68927c9
feat: add support for tensor parallel flow using accelerate
kmehant Oct 15, 2024
87bebaf
fix: add tp degree to env variable
kmehant Dec 13, 2024
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
8 changes: 7 additions & 1 deletion src/transformers/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -233,6 +233,7 @@
from accelerate.utils import (
DistributedDataParallelKwargs,
DistributedType,
TorchTensorParallelPlugin,
load_fsdp_model,
load_fsdp_optimizer,
save_fsdp_model,
Expand Down Expand Up @@ -5055,6 +5056,11 @@ def create_accelerator_and_postprocess(self):
args["dataloader_config"] = dataloader_config
else:
args.update(accelerator_config)
# tp is initialized at Accelerator init phase so
# args should be prepared here
if self.args.tp_size > 1:
self.is_tp_enabled = True
args["torch_tp_plugin"] = TorchTensorParallelPlugin(tp_size=self.args.tp_size)

# create accelerator object
self.accelerator = Accelerator(**args)
Expand All @@ -5069,7 +5075,7 @@ def create_accelerator_and_postprocess(self):
# deepspeed and accelerate flags covering both trainer args and accelerate launcher
self.is_deepspeed_enabled = getattr(self.accelerator.state, "deepspeed_plugin", None) is not None
self.is_fsdp_enabled = getattr(self.accelerator.state, "fsdp_plugin", None) is not None

self.is_tp_enabled = getattr(self.accelerator.state, "tp_plugin", None) is not None
# post accelerator creation setup
if self.is_fsdp_enabled:
fsdp_plugin = self.accelerator.state.fsdp_plugin
Expand Down
17 changes: 16 additions & 1 deletion src/transformers/training_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -565,7 +565,9 @@ class TrainingArguments:
Will use gradient checkpointing over each nested XLA FSDP wrapped layer. This setting can only be
used when the xla flag is set to true, and an auto wrapping policy is specified through
fsdp_min_num_params or fsdp_transformer_layer_cls_to_wrap.

tp_size (`int`, *optional*):
Use tp_size to enable PyTorch tensor parallelism. Set a value greater than 1 to activate TP. The same is
used to prepare device mesh internally.
deepspeed (`str` or `dict`, *optional*):
Use [Deepspeed](https://github.com/microsoft/deepspeed). This is an experimental feature and its API may
evolve in the future. The value is either the location of DeepSpeed json config file (e.g.,
Expand Down Expand Up @@ -1239,6 +1241,16 @@ class TrainingArguments:
)
},
)
tp_size: Optional[int] = field(
default=0,
metadata={
"help": (
"Use tp_size to enable pytorch tensor parallelism."
"Set a value greater than 1 to activate TP."
"The same is used to prepare device mesh internally."
)
},
)
fsdp_transformer_layer_cls_to_wrap: Optional[str] = field(
default=None,
metadata={
Expand Down Expand Up @@ -1964,6 +1976,9 @@ def __post_init__(self):
if self.fsdp_config["xla_fsdp_grad_ckpt"]:
warnings.warn("`--xla_fsdp_grad_ckpt` is useful only when `--xla` is set to true.")

if self.tp_size > 1:
os.environ["ACCELERATE_USE_TP"] = "true"
os.environ["TP_SIZE"] = self.tp_size
# accelerate integration for FSDP
if len(self.fsdp) > 0 and not self.fsdp_config["xla"]:
os.environ["ACCELERATE_USE_FSDP"] = "true"
Expand Down