Training Module CLI
The NuPIC Training Module is started using the nupic/nupic_training.sh
script.
To use this script, make sure it is made executable using chmod +x nupic/nupic_training.sh
, then use the following syntax:
./nupic_training.sh [COMMANDS] [OPTIONS]
Commands
Command | Description |
---|---|
start | Starts a new NuPIC Training Module. |
stop | Stops the running NuPIC Training Module. |
remove | Removes existing container of NuPIC Training Module. |
update_image | Stops and removes the running NuPIC Training Module container, updates and loads an image for NuPIC Training Module. This is useful for updating the Training Module from a local image, without having to entirely reinstall NuPIC. To be used in conjunction with --image_path . |
Options
Option | Value | Description |
---|---|---|
--cpus | CPU indices | Defines the explicit CPUs which NuPIC Training Module will run on. If not specified, NuPIC Training Module runs on all available CPU cores. This can be useful when running NuPIC Training Module concurrently with other workloads on the same machine. |
--gpus | GPU indices | Defines the explicit GPUs which NuPIC Training Module will run on. If not specified, NuPIC Training Module runs on all available GPU cards. Alternatively, specify an invalid GPU index (e.g., 999 ) to prevent the Training Module from using any GPUs your machine. This forces the Training Module to only use the CPU. |
--port | Port number | The port NuPIC Training Module listens to. If not specified, port 8321 will be used by default. |
--image_path | File path | The path to the docker image of NuPIC Training Module to be loaded. This option is only used with update_image command. |
Updated 4 months ago