Skip to content

[TorchAcc] Add USE_TORCH_XLA=0 flag for native swift scripts #1221

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

Merged
merged 2 commits into from
Jun 25, 2024
Merged

[TorchAcc] Add USE_TORCH_XLA=0 flag for native swift scripts #1221

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
f88741d
add USE_TORCH_XLA=0 flag
anw90 Jun 25, 2024
858bd35
remove trim graph
anw90 Jun 25, 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
3 changes: 2 additions & 1 deletion examples/pytorch/llm/scripts/torchacc/baichuan2_13b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,8 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \

export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
Expand Down
4 changes: 2 additions & 2 deletions examples/pytorch/llm/scripts/torchacc/chatglm3_6b/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,8 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
# MASTER_PORT=12356 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
swift sft \
Expand Down
2 changes: 1 addition & 1 deletion examples/pytorch/llm/scripts/torchacc/llama2_13b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
Expand Down
1 change: 0 additions & 1 deletion examples/pytorch/llm/scripts/torchacc/llama3_8b_instruct/acc_lora_fsdp_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,6 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.
export USE_TORCHACC=1
export TORCHACC_TRIM_GRAPH=1
export XLA_IR_SHAPE_CACHE_SIZE=100000000
export XLA_ALLOCATOR_FRACTION=0.95
export XLA_EXPERIMENTAL=nonzero:masked_select
Expand Down
2 changes: 1 addition & 1 deletion examples/pytorch/llm/scripts/torchacc/llama3_8b_instruct/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
Expand Down
2 changes: 1 addition & 1 deletion examples/pytorch/llm/scripts/torchacc/qwen1half_14b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1 \
Expand Down
2 changes: 1 addition & 1 deletion examples/pytorch/llm/scripts/torchacc/qwen1half_32b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
Expand Down
6 changes: 3 additions & 3 deletions examples/pytorch/llm/scripts/torchacc/qwen_72b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,10 +2,10 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=1 \
CUDA_VISIBLE_DEVICES=7 \
NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
--model_id_or_path qwen/Qwen-72B-Chat \
--dataset codefuse-python-en \
Expand Down
2 changes: 1 addition & 1 deletion examples/pytorch/llm/scripts/torchacc/yi_34b_chat/swift_lora_sft.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
# 80GB GPU memory
# Note: TorchAcc is currently only available internally.

# MASTER_ADDR=127.0.0.1 \
export USE_TORCH_XLA=0

NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
Expand Down
Loading