grid logs
Get logs when running experiments
To view logs for an experiment use:
1
grid logs the-experiment-name
Copied!
Logs show two sections, build logs and stdout logs.
1
# build logs
2
[build] [2021-04-20T01:42:31.045408+00:00] Task experiments-cd13d646-cf31-486a-a78f-2e4a825e27dd created.
3
...
4
5
# stdout logs
6
[stdout] [2021-04-20T01:47:59.723077+00:00] Restoring data from /gridai/project/lightning_logs to .
7
[stdout] [2021-04-20T01:48:06.702111+00:00] --------------------------------------------------
8
[stdout] [2021-04-20T01:48:06.702137+00:00] GPUS: There are 1 GPUs on this machine
9
[stdout] [2021-04-20T01:48:06.702144+00:00] --------------------------------------------------
10
[stdout] [2021-04-20T01:48:06.702150+00:00] PARAMS: I want to eat: 3 pear
11
[stdout] [2021-04-20T01:48:06.702184+00:00] --------------------------------------------------
12
[stdout] [2021-04-20T01:48:06.702192+00:00] i can run any ML library like numpy, pytorch lightning, sklearn pytorch, keras, tensorflow
13
[stdout] [2021-04-20T01:48:06.702199+00:00] torch: tensor([0.6885]) numpy [0.9643539]
Copied!

Build logs

Sometimes an experiment may fail because a dependency install can be wrong, or something happens during the environment build.
These kinds of failures can be found in the build logs.
1
[build] [2021-04-20T01:42:31.045408+00:00] Task experiments-cd13d646-cf31-486a-a78f-2e4a825e27dd created.
Copied!

Experiment logs

For outputs of the file, look at the stdout logs. If your file throws an exception, that will appear here.
1
# stdout logs
2
[stdout] [2021-04-20T01:47:59.723077+00:00] Restoring data from /gridai/project/lightning_logs to .
3
[stdout] [2021-04-20T01:48:06.702111+00:00] --------------------------------------------------
4
[stdout] [2021-04-20T01:48:06.702137+00:00] GPUS: There are 1 GPUs on this machine
5
[stdout] [2021-04-20T01:48:06.702144+00:00] --------------------------------------------------
6
[stdout] [2021-04-20T01:48:06.702150+00:00] PARAMS: I want to eat: 3 pear
7
[stdout] [2021-04-20T01:48:06.702184+00:00] --------------------------------------------------
8
[stdout] [2021-04-20T01:48:06.702192+00:00] i can run any ML library like numpy, pytorch lightning, sklearn pytorch, keras, tensorflow
9
[stdout] [2021-04-20T01:48:06.702199+00:00] torch: tensor([0.6885]) numpy [0.9643539]
Copied!

Command arguments

--n_lines

Number of log lines (from the end) to return.

--page

When an experiment is finished, logs are given in pages. This specificies which page to fetch from archived logs.

--max_lines

Maximum number of lines to print in terminal

--use_pager

Allows interactive / scrollable logs
Last modified 7mo ago