kerasvideo's Introduction

Getting Started

You'll need to run this with a Docker capable host computer, installation support is provided on Ubuntu.

install-docker.sh -- this will get Docker up and running

Now you can get a container ready with:

docker build --tag keras-cpu .
docker run -p 8888:8888 --volume $(pwd):/src keras-cpu

GPU Support

gpu/install-nvidia-docker.sh -- this requires you have a NVIDIA graphics card as well as the current driver.

If all is well at this point, you will see an inventory of your graphics cards such as:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.77                 Driver Version: 390.77                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  TITAN V             Off  | 00000000:03:00.0 Off |                  N/A |
| 37%   53C    P0    38W / 250W |      0MiB / 12066MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  TITAN V             Off  | 00000000:04:00.0 Off |                  N/A |
| 33%   48C    P0    37W / 250W |      0MiB / 12066MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

Now you can get a container ready with:

docker build --tag keras-gpu ./gpu
nvidia-docker run -p 8888:8888 --volume $(pwd):/src keras-gpu

kerasvideo's People

Contributors

Stargazers

Watchers

kerasvideo's Issues

running docker build - hangs on 'useradd' step - actually its the chown /opt/conda step

Not sure why but I've tried this on 2 different windows for docker latest images. The step where useradd occurs, which chains to chown /opt/conda -R... - that process eventually blocks and goes into an uninterruptable sleep state D

The following I got by docker exec into the build container at this state to see it's state.

USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.0  0.0   4500    72 ?        Ss   05:32   0:00 /bin/sh -c useradd -m -s /bin/bash -N -u $NB_UID $NB_USER &&     mkdirroot        12  3.7  0.0  12960   852 ?        D    05:32   0:45 chown keras /opt/conda -R
root        13  0.0  0.0  18248  2832 pts/0    Ss   05:43   0:00 bash
root        99  0.0  0.0  34420  2820 pts/0    R+   05:52   0:00 ps aux

The step shown in the console from build is

creating default environment...
using -f (force) option
installation finished.
Removing intermediate container fe2853a4e525
 ---> 2a5b9f8d3564
Step 8/14 : RUN useradd -m -s /bin/bash -N -u $NB_UID $NB_USER &&     mkdir -p $CONDA_DIR &&     chown keras $CONDA_DIR -R &&     mkdir -p /src &&     chown keras /src
 ---> Running in cdc73aad980d

Making it all run as root, removing the chown step, works fine

From your Pack book p21

Hi, i'm page 21 from your book (deep learning images with tensorflow)
I have a something wrong, i can't import keras in my IPYTHON. Why? Is the Ipython linked with the docker ?