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.

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

Validation Accuracy different from paper #20

Open
soonchangAI opened this issue May 30, 2022 · 0 comments
Open

Validation Accuracy different from paper #20

soonchangAI opened this issue May 30, 2022 · 0 comments

Comments

@soonchangAI
Copy link

soonchangAI commented May 30, 2022

Hi, the validation accuracy I calculated for the fine-tuned models are different from the paper.
Command:

python -m torch.distributed.launch --nproc_per_node 2  tools/run.py --tasks vqa --datasets m4c_textvqa --model m4c_split \
--config $config \
--save_dir $folder \
--run_type val \
--resume_file $finetuned_model \
training_parameters.distributed True 

I observed changing the batch size results in different values.

Val accuracy for batch size = 32 Val acc for batch size = 128 In paper
TextVQA TAP (base) 49.87 49.53 49.91
TextVQA TAP (additional data) 54.31 54.13 54.71
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant