When running the convnet.py, I encountered the following error:
“1.1 (0.00%)...cannot allocate memory for thread-local data: ABORT”.
I have no idea about this.
By the way, I used cuda-convnet2 and the codes are running on Red Hat.
Original issue reported on code.google.com by kgl.p...@gmail.com on 30 Nov 2014 at 4:21
Original issue reported on code.google.com by
kgl.p...@gmail.com
on 30 Nov 2014 at 4:21