You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docker/README.md
+12-5
Original file line number
Diff line number
Diff line change
@@ -15,22 +15,28 @@ Refer to [Install for GPU](../docs/install/install_for_gpu.md) and [Install for
15
15
16
16
## Binaries Preparation
17
17
18
-
Download and copy Intel® Extension for TensorFlow* wheel into ./models/binaries directory.
18
+
Download and copy Intel® Extension for TensorFlow* wheel into ./models/binaries directory.
19
19
20
20
```
21
21
mkdir ./models/binaries
22
22
```
23
23
24
+
To use Intel® Optimization for Horovod* with the Intel® oneAPI Collective Communications Library (oneCCL), copy Horovod wheel into ./models/horovod as well.
25
+
26
+
```bash
27
+
mkdir ./models/horovod
28
+
```
29
+
24
30
## Usage of Docker Container
25
31
### I. Customize build script
26
32
[build.sh](./build.sh) is provided as docker container build script. While OS version and some software version (such as Python and TensorFlow) is hard coded inside the script. If you prefer to use newer or later version, you can edit this script.
27
33
28
-
For example, to build docker container with Python 3.9 and TensorFlow 2.10 on Ubuntu 20.04 layer, update [build.sh](./build.sh) as below.
34
+
For example, to build docker container with Python 3.9 and TensorFlow 2.11 on Ubuntu 20.04 layer, update [build.sh](./build.sh) as below.
Run following commands to start docker container. You can use `-v` option to mount your local directory into container. To make GPU available in the container, attach the GPU to the container using `--device /dev/dri` option and run the container:
0 commit comments