1.2.0

• Added --meter CLI option to show graphical utilization meters (#25, thanks @rit628 )

1.1.0

• Added CLI argument --id to filter the processed GPUs (it is passed to nvidia-smi).
• Changed some error text to be printed to stderr instead of on stdout.
• The program now exits with the same error code as nvidia-smi if its call failed.

1.0.7

Added possibility to filter processes by user(s).

1.0.6

Fixed:

• coloring of the output.

1.0.5

Maintenance release

1.0.4

Fixed:

• display in case there are processes with PIDs >= 1M

1.0.3

Fixed outputting commands longer than 100 characters.

1.0.2

Release on PyPi.