Skip to content

[data] feat: TransferQueue - Add zero-copy serialization support & usage improvement #4429

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
wuxibin89 merged 9 commits into from
Dec 15, 2025
Merged

[data] feat: TransferQueue - Add zero-copy serialization support & usage improvement #4429

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
b526544
update TQ version
0oshowero0 Dec 5, 2025
a000162
add TQ_ZERO_COPY_SERIALIZATION env
0oshowero0 Dec 5, 2025
b68ac8e
del requirements_transferqueue.txt
0oshowero0 Dec 8, 2025
319396d
fix
0oshowero0 Dec 13, 2025
1c37c89
add logger & optimize batchmeta usage
0oshowero0 Dec 15, 2025
3324c4a
fix
0oshowero0 Dec 15, 2025
1ce2e9a
fix
0oshowero0 Dec 15, 2025
48c72cb
fix
0oshowero0 Dec 15, 2025
87cc069
fix
0oshowero0 Dec 15, 2025
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
10 changes: 6 additions & 4 deletions .github/workflows/e2e_transferqueue.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,13 +124,14 @@ jobs:
run: |
pip3 install --no-deps -e .[test,gpu]
pip3 install transformers==$TRANSFORMERS_VERSION
pip3 install -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple TransferQueue==0.1.2.dev0
pip3 install -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple TransferQueue==0.1.4.dev1
- name: Prepare GSM8K dataset
run: |
python3 examples/data_preprocess/gsm8k.py --local_dataset_path ${HOME}/models/hf_data/gsm8k
- name: Running the E2E test with TransferQueue (FSDP)
- name: Running the E2E test with TransferQueue (FSDP), enable zero copy serialization
run: |
ray stop --force
export TQ_ZERO_COPY_SERIALIZATION=True
bash tests/special_e2e/run_transferqueue.sh

# Test Megatron strategy
Expand All @@ -153,13 +154,14 @@ jobs:
run: |
pip3 install --no-deps -e .[test,gpu]
pip3 install transformers==$TRANSFORMERS_VERSION
pip3 install -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple TransferQueue==0.1.2.dev0
pip3 install -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple TransferQueue==0.1.4.dev1
- name: Prepare GSM8K dataset
run: |
python3 examples/data_preprocess/gsm8k.py --local_dataset_path ${HOME}/models/hf_data/gsm8k
- name: Running the E2E test with TransferQueue (Megatron)
- name: Running the E2E test with TransferQueue (Megatron), disable zero copy serialization
run: |
ray stop --force
export TQ_ZERO_COPY_SERIALIZATION=False
bash tests/special_e2e/run_transferqueue.sh

cleanup:
Expand Down
Loading
Loading