nqanh / affordance-net

AffordanceNet - Multiclass Instance Segmentation Framework - ICRA 2018
Other
121 stars 54 forks source link

Question on memory error #2

Open Tetsujinfr opened 6 years ago

Tetsujinfr commented 6 years ago

Hi,

I got often memory issue on images, is it due to my 980Ti having "only" 6GB of RAM you think?

For instance I got this message below for the attached image:

Current idx:  0  /  1
Current img:  6893.jpg
F1006 07:28:27.337713 13705 syncedmem.cpp:56] Check failed: error == cudaSuccess (2 vs. 0)  out of memory

6893

Tets

nqanh commented 6 years ago

huh, you need around 6.xx GB ram for testing this full AffordanceNet. I'll release a smaller version soon. You can also use CPU to test some images. It's just slower (takes around 20s/image on my PC). But you will need around 8GB RAM for this.

superchenyan commented 6 years ago

Maybe there are some problems in the test code. The author's model is ok. I train the model based on the author provided dataset, I find it randomly print Check failed: error == cudaSuccess (2 vs. 0) while test different iter models, especially the small iter model, such as iter_20.caffemodel.