Samplers¶
-
class
pytorch_wrapper.samplers.
OrderedBatchWiseRandomSampler
(data_source, get_order_value_callable, batch_size, seed=1234)¶ Bases:
sphinx.ext.autodoc.importer._MockObject
Semi-randomly samples indexes from a dataset ensuring that the corresponding examples will have similar values. Values are returned by a callable.
Parameters: - data_source – a data source (usually a dataset object).
- get_order_value_callable – a callable that takes as input the example’s index and returns the ordering value.
- batch_size – the batch size.
- seed – the initial seed.
-
class
pytorch_wrapper.samplers.
OrderedSequentialSampler
(data_source, get_order_value_callable)¶ Bases:
sphinx.ext.autodoc.importer._MockObject
Samples elements from a dataset ordered by a value returned by a callable for each example.
Parameters: - data_source – a data source (usually a dataset object).
- get_order_value_callable – a callable that takes as input the example’s index and returns the ordering value.
-
class
pytorch_wrapper.samplers.
SubsetOrderedBatchWiseRandomSampler
(indexes, get_order_value_callable, batch_size, seed=1234)¶ Bases:
sphinx.ext.autodoc.importer._MockObject
Semi-randomly samples indexes from a list ensuring that the corresponding examples will have similar values. Values are returned by a callable.
Parameters: - indexes – a list of indexes.
- get_order_value_callable – a callable that takes as input the example’s index and returns the ordering value.
- batch_size – the batch size.
- seed – the initial seed.
-
class
pytorch_wrapper.samplers.
SubsetOrderedSequentialSampler
(indexes, get_order_value_callable)¶ Bases:
sphinx.ext.autodoc.importer._MockObject
Samples elements from a list of indexes ordered by a value returned by a callable for each example.
Parameters: - indexes – a list of indexes.
- get_order_value_callable – a callable that takes as input the example’s index and returns the ordering value.
-
class
pytorch_wrapper.samplers.
SubsetSequentialSampler
(indexes)¶ Bases:
sphinx.ext.autodoc.importer._MockObject
Samples elements sequentially based on a list of indexes.
Parameters: indexes – a list of indexes.