I am trying to initialize a CNN and then put it on my GPU for training. When I put it on GPU I get the error: (CUDA error: out of memory). I have run similar networks with no such problems. This is the only thing in cuda as I have not loaded any images as of yet. Any ideas as to what is going wrong?
I am using pytorch version 0.4.1 on a GTX 1070ti 8GB.
| NVIDIA-SMI 410.104 Driver Version: 410.104 CUDA Version: 10.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 107... Off | 00000000:01:00.0 On | N/A |
| 0% 43C P2 39W / 180W | 8024MiB / 8111MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1129 G /usr/lib/xorg/Xorg 36MiB |
| 0 1164 G /usr/bin/gnome-shell 57MiB |
| 0 1415 G /usr/lib/xorg/Xorg 200MiB |
| 0 1548 G /usr/bin/gnome-shell 90MiB |
| 0 6323 C /usr/bin/python3 525MiB |
| 0 9521 C /usr/bin/python3 1827MiB |
| 0 18821 C /usr/bin/python3 4883MiB |
| 0 27137 G ...uest-channel-token=16389326112703159917 45MiB |
| 0 29161 C /usr/bin/python3 355MiB |
I have tried reducing the size of the linear layers with no luck.
net = piccnn()
net.to(device)
This issue happened to me once when a GPU driver was out of date. My GPU was a 1070 4 gig. I'd recommend a reinstall of drivers and restart.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.