search

ying09 / TextFuseNet

A PyTorch implementation of "TextFuseNet: Scene Text Detection with Richer Fused Features".
MIT License
476 stars 123 forks source link

Use pre-trained model for prediction without GPU #49

Closed rbgt closed 3 years ago

rbgt commented 3 years ago

Hi, I'd like to try the TextFuseNet architecture without training on new data but only to assess the performance of the model, is it possible to do it without GPU ? I've followed the step by step installation guide and placed the detection model in a created folder according to the python file but when running the demo, it needs to have a GPU.

Is there something missing here ?

rbgt commented 3 years ago

Alright just found out it's possible to run prediction on CPU by setting "cpu" instead of "cuda" at one line in the "detectron2/config/default.py" file :

from : _C.MODEL.DEVICE = "cuda" switch to : _C.MODEL.DEVICE = "cpu"

chenf99 commented 3 years ago

@rbgt Hi, did you run prediction on CPU successfully? I have modified _C.MODEL.DEVICE = "cuda" to _C.MODEL.DEVICE = "cpu"，but it shows Segmentation fault (core dumped).