razorx89 / roco-dataset

Radiology Objects in COntext (ROCO): A Multimodal Image Dataset
175 stars 19 forks source link

I don't know how to complete the download? #15

Closed languandong closed 3 months ago

languandong commented 9 months ago

I run the script "python scripts/fetch.py". But it downloaded as 9.895%. image So I immediately re-executed the script. image Should I take -d at the end of this instruction?

saviola777 commented 9 months ago

Did the script abort after the errors in the first screenshot? Usually it should just continue even if some images are no longer available (which is to be expected after 5+ years). If you pass -d it will delete the images you have downloaded so far.

For an alternative dataset, you might want to check out ROCOv2.

languandong commented 9 months ago

Did the script abort after the errors in the first screenshot? Usually it should just continue even if some images are no longer available (which is to be expected after 5+ years). If you pass -d it will delete the images you have downloaded so far.

For an alternative dataset, you might want to check out ROCOv2.

This V2's form is so convenient. I like it! Maybe V1 can also be organized like this

biegekekeke commented 9 months ago

Did the script abort after the errors in the first screenshot? Usually it should just continue even if some images are no longer available (which is to be expected after 5+ years). If you pass -d it will delete the images you have downloaded so far.

For an alternative dataset, you might want to check out ROCOv2.

Why doesn't v2 have keywords anymore?

saviola777 commented 9 months ago

Good question! In the end you can probably generate keywords using the concept CUIs plus the concept names. Maybe we will add keywords at a later stage, thanks for the suggestion.

biegekekeke commented 9 months ago

Good question! In the end you can probably generate keywords using the concept CUIs plus the concept names. Maybe we will add keywords at a later stage, thanks for the suggestion.

Oh, yes, I have incorporated CUIs along with concept names to generate keywords. Thank you for your help.