Skip to content
New issue

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

#

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

Already on GitHub? # to your account

Tune VLM configs for SmolVLM and Qwen2-VL #1307

Merged
merged 2 commits into from
Jan 29, 2025
Merged

Tune VLM configs for SmolVLM and Qwen2-VL #1307

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
5a93fa8
save
xrdaukar Jan 29, 2025
4a77711
save
xrdaukar Jan 29, 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
3 changes: 1 addition & 2 deletions configs/recipes/vision/qwen2_vl_2b/sft/train.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@ data:
transform_num_workers: "auto"
dataset_kwargs:
processor_name: "Qwen/Qwen2-VL-2B-Instruct"
limit: 4096
# limit: 4096 # Uncomment to limit dataset size!
return_tensors: True
# - dataset_name: "HuggingFaceH4/llava-instruct-mix-vsft"
# split: "train"
Expand All @@ -32,7 +32,6 @@ data:
# transform_num_workers: "auto"
# dataset_kwargs:
# processor_name: "Qwen/Qwen2-VL-2B-Instruct"
# limit: 8192
# return_tensors: True

training:
Expand Down
2 changes: 1 addition & 1 deletion configs/recipes/vision/smolvlm/sft/gcp_job.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@ setup: |
pip install uv && uv pip install '.[gpu]' hf_transfer
# Install model from HF Hub. This tool increases download speed compared to
# downloading the model during training.
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download HuggingFaceTB/SmolVLM-Instruct
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download HuggingFaceTB/SmolVLM-Instruct --exclude "onnx/*" "runs/*"

pip install -U "transformers>=4.46,<4.47"
pip install -U flash-attn --no-build-isolation
Expand Down
2 changes: 1 addition & 1 deletion configs/recipes/vision/smolvlm/sft/train.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ data:
transform_num_workers: "auto"
dataset_kwargs:
processor_name: "HuggingFaceTB/SmolVLM-Instruct"
limit: 4096
# limit: 4096 # Uncomment to limit dataset size!
return_tensors: True

training:
Expand Down