if context[0] in ['img', 'both']: assert np.array(cxt_modality_label).size == vis_feats.size(0) == vis_pe.size(0) AssertionError
Opened this issue · 2 comments
MuHeDing commented
I have this bug and want to seek for solutions.
zdxdsw commented
Hi Muhe:
Can you please print out cxt_modality_label.size(), vis_feats.size() and vis_pe.size()? Basically this line is just confirming that the dataloader returns same-length variables.
MuHeDing commented
Thank you for your answers!I have run both single-gpu and multi-gpu modes, and found that there is no problem with the single gpu, but the multi-gpu error is reported, which is the error I proposed. This is the screenshot of the dimension output, please take a look.
gpu=1
gpu=4
…------------------ 原始邮件 ------------------
发件人: ***@***.***>;
发送时间: 2022年12月20日(星期二) 凌晨3:51
收件人: ***@***.***>;
抄送: ***@***.***>; ***@***.***>;
主题: Re: [WebQnA/WebQA_Baseline] if context[0] in ['img', 'both']: assert np.array(cxt_modality_label).size == vis_feats.size(0) == vis_pe.size(0) AssertionError (Issue #9)
Hi Muhe:
Can you please print out cxt_modality_label.size(), vis_feats.size() and vis_pe.size()? Basically this line is just confirming that the dataloader returns same-length variables.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>