osl_dynamics.config_api
#
Config API.
Specify a pipeline using a config, e.g. to train a TDE-HMM:
config = ```
load_data:
inputs: training_data
prepare:
tde_pca: {n_embeddings: 15, n_pca_components: 80}
standardize: {}
train_hmm:
config_kwargs:
n_states: 8
learn_means: False
learn_covariances: True
```
and run with:
run_pipeline(config, output_dir="results")
See the toolbox paper examples for scripts that use the config API:
Note
The config API can be used via the command line with:
% osl-dynamics <config-file> <output-directory>
where
<config-file>
is a yaml file containing the config.<output-directory>
is the output directory.
Optionally, you can specify a particular GPU to use with:
% osl-dynamics <config-file> <output-directory> --restrict <restrict>
where <restrict>
is an integer specifying the GPU number. E.g. if you
would just like to use the first GPU, you can pass:
% osl-dynamics <config-file> <output-directory> --restrict 0
Remember you need to activate the osld
conda environment to use the
command line interface.