VikParuchuri / marker

Convert PDF to markdown quickly with high accuracy
https://www.datalab.to
GNU General Public License v3.0
15.97k stars 883 forks source link

Script stops running after a while with big file sizes (~80-100MB). #230

Open joeamroo opened 1 month ago

joeamroo commented 1 month ago

Hi! So I'm running a script that would take 40ish files using 16 workers. Problem is it stops working after a while

image

It almost always stops after the last few files get their reading order done.

joeamroo commented 1 month ago

@VikParuchuri Do you know why that might be happening? left it running overnight and it got stuck for some reason, I have an A100 with 5.5GB VRAM_PER_TASK.

joeamroo commented 1 month ago

It gets stuck after finishing Detecting bboxes and the processes just rest at those memory sizes.

Screenshot 2024-07-17 at 10 38 22 AM Screenshot 2024-07-17 at 10 35 55 AM
joeamroo commented 1 month ago

205 Might be related.