Jue
Jue
I faced exactly same issue.
@MBakirWB I was running a personal local instance to try wandb out. https://github.com/wandb/local/issues/64#issuecomment-1106030481
I met same issue trying initializing `reader = easyocr.Reader(["en", "de"], detect_network="dbnet18")` ``` File ~/.virtualenvs/py3106/lib/python3.10/site-packages/easyocr/DBNet/assets/ops/dcn/functions/deform_conv.py:5, in 2 from torch.autograd import Function 3 from torch.nn.modules.utils import _pair ----> 5 from .. import...
I've tried various combinations with * python 3.7/3.8/3.9/3.10, * torch 1.7/1.9/1.10/1.12, * cuda 10.1/11.3/11.6/11.7, * ubuntu 20.04/22.04, * via `pip instal easyocr`/with `python compile_dbnet_dcn.py`. Unfortunately all of them failed at...
Update: I got it work. The key is: * nvcc is necessary, so cuda runtime image as base image is not enough. It must be cuda-devel. * env var CUDA_HOME...
> 思路1： 合并 有重合的检测框，判断两个框在同一水平线上并且有重合，就将两个检测框合并为一个 我现在的workaround是在整个检测+识别结束后，扔掉跟别的框重合的单个字符（假设所有字符+空白都等宽分配框的宽度）。感觉要是能实现这个思路1更好。
奇怪，似乎是2.8.1的问题，我退到paddleocr 2.7.3就没有这个显存泄漏的问题了
Any updates or maybe workarounds? I am facing same issue and I can't even cancel the queue-ed dependabot job...