get_model_descr
get_model_descr(model, source='kipoi')
Get model description
Arguments
- model: model's relative path/name in the source. 2nd column in the
kipoi.list_models()
pd.DataFrame`. - source: Model source. 1st column in the
kipoi.list_models()
pd.DataFrame
.
get_model
get_model(model, source='kipoi', with_dataloader=True, **kwargs)
Load the model
from source
, as well as the default dataloder to model.default_dataloder.
Arguments
- model (str): model name
- source (str): source name
- with_dataloader (bool): if True, the default dataloader is
loaded to
model.default_dataloader
and the pipeline atmodel.pipeline
enabled.
Returns
- Instance of class inheriting from
kipoi.models.BaseModel
(likekipoi.models.KerasModel
) decorated with additional attributes.
Methods
- predict_on_batch(x): Make model predictions given a batch of data
x
Appended attributes
- type (
str
): model type (class name) - args (
dict
): model args used to instantiate the model class - info (
kipoi.specs.Info
): information about the author (etc) - schema (
kipoi.specs.ModelSchema
): information about the input/outputdata modalities - dependencies (
kipoi.specs.Dependencies
): class specifying the dependencies. (implementsinstall
method for running the installation) - default_dataloader (class inheriting from
kipoi.data.BaseDataLoader
): default dataloader. None ifwith_dataloader=False
was used. - name (
str
): model name - source (
str
): model source - source_dir (
str
): local path to model source storage - writers (dict): dictionary of arguments for writers
- pipeline (
kipoi.pipeline.Pipeline
): handle to aPipeline
object