python gym-DQN.py => RuntimeError: torch was compiled without numpy support

[TheRobotStudio]

Hello, After successfully build, with a fresh install of a Jetpack 3.0 on a TX2, when I try to run gym-DQN.py, I got this error:

$ python gym-DQN.py 
[2017-07-21 14:08:36,674] Making new env: CartPole-v0
Traceback (most recent call last):
  File "gym-DQN.py", line 280, in <module>
    plt.imshow(get_screen().cpu().squeeze(0).permute(1, 2, 0).numpy(),
  File "gym-DQN.py", line 274, in get_screen
    screen = torch.from_numpy(screen)
RuntimeError: torch was compiled without numpy support
/usr/local/lib/python2.7/dist-packages/matplotlib/backends/backend_gtk3.py:215: Warning: Source ID 8 was not found when attempting to remove it
  GLib.source_remove(self._idle_draw_id)

Also I had to install matplotlib for python: $ sudo pip install --upgrade matplotlib But it seems there is a problem with numpy. Any idea on how to solve this issue ? Is there a list of pyhton modules that must be installed before the build ? Thanks!

[dusty-nv]

Hi, sorry for the problem, perhaps the pre-install command to install numpy is missing (only used in the standalone examples). I guess my system already had it from before.

Can you try installing it like this?

sudo pip install numpy

Then rm build and make a new build.

Hope that helps, Dustin

On Jul 21, 2017 8:20 AM, Cyril notifications@github.com wrote:

Hello, After successfully build, with a fresh install of a Jetpack 3.0 on a TX2, when I try to run gym-DQN.py, I got this error:

$ python gym-DQN.py [2017-07-21 14:08:36,674] Making new env: CartPole-v0 Traceback (most recent call last): File "gym-DQN.py", line 280, in plt.imshow(get_screen().cpu().squeeze(0).permute(1, 2, 0).numpy(), File "gym-DQN.py", line 274, in get_screen screen = torch.from_numpy(screen) RuntimeError: torch was compiled without numpy support /usr/local/lib/python2.7/dist-packages/matplotlib/backends/backend_gtk3.py:215: Warning: Source ID 8 was not found when attempting to remove it GLib.source_remove(self._idle_draw_id)

Also I had to install matplotlib for python: $ sudo pip install --upgrade matplotlib But it seems there is a problem with numpy. Any idea on how to solve this issue ? Is there a list of pyhton modules that must be installed before the build ? Thanks!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/dusty-nv/jetson-reinforcement/issues/6, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AOpDK_nNbrLIJejILsDx6v8qH7vi8hIdks5sQJeLgaJpZM4OfVTV.

---

This email message is for the sole use of the intended recipient(s) and may contain confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.

[TheRobotStudio]

Thanks Dustin, After doing that, I got the error message: TypeError: Couldn't find foreign struct converter for 'cairo.Context' I could solve this installing this, in case some people has this issue link: sudo apt-get install python-gi-cairo And now the demo is working!

[dusty-nv]

OK thank you Cyril, I will add these commands to the pre-build script.

On Jul 21, 2017 10:45 AM, Cyril notifications@github.com wrote:

Thanks Dustin, After doing that, I got the error message: TypeError: Couldn't find foreign struct converter for 'cairo.Context' I could solve this installing this, in case some people has this issue linkhttps://github.com/rbgirshick/py-faster-rcnn/issues/221: sudo apt-get install python-gi-cairo And now the demo is working!

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/dusty-nv/jetson-reinforcement/issues/6#issuecomment-317020310, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AOpDK_KGB5mJydKohpmxfV1oQsmwMfyYks5sQLlfgaJpZM4OfVTV.

---

This email message is for the sole use of the intended recipient(s) and may contain confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.