Command reference¶
This section lists every neuroatlas command and its options, one page per
group of commands. Run neuroatlas <command> --help to see the same text in
the terminal. The user guide shows how the commands fit
together.
Global options¶
These options work with every command, before or after the command name.
Usage: neuroatlas [-v] [--log FILE] [--online] <command> [options]
| Option | Description |
|---|---|
-v, --verbose |
Show more detail, library log lines and full tracebacks. |
--log FILE |
Also write all output, log lines and tracebacks to FILE. |
--online |
Allow downloads. Only the commands that download are online by default. |
-V, --version |
Print the version, and the commit when run from a git clone. |
Every command exits with status 0 when it succeeds, 1 when a run fails and 2 when the command line is wrong.
Commands¶
| Group | Command | Description |
|---|---|---|
| Setup | config |
Set where data, caches, results and model weights are stored. |
| Explore | list |
List the benchmarks, datasets, models, aliases or tasks. |
show |
Describe a benchmark and print the commands it runs. | |
| Data and weights | data |
Check, download and prepare datasets. |
models |
Check and download model weights. | |
| Run | check |
Test each dataset and model on one batch before a long run. |
run |
Run a benchmark on this machine. | |
submit |
Write HTCondor or SLURM jobs for a benchmark. | |
status |
Show the state of the jobs that submit wrote. |
|
| Results | results |
Summarise a benchmark's results. |
rescore |
Recompute the metrics from saved test predictions. | |
| Individual steps | embed |
Extract embeddings of one dataset with frozen models. |
probe |
Fit probes on extracted embeddings. | |
hypnogram |
Compute hypnograms and sleep features from sleep staging results. |