gigl.src.inference.v1.lib.base_inferencer#

Attributes#

`T`
`logger`

Classes#

`BaseInferencer`	The Protocol that you need to implement for your inferencer to function with automated Inference
`InferBatchResults`
`NodeAnchorBasedLinkPredictionBaseInferencer`	The protocol that you need to implement for inderence for Node Anchor Based Link Prediction tasks in tabularized mode.
`SupervisedNodeClassificationBaseInferencer`	The protocol that you need to implement for inference for Supervised Node Classification tasks in tabularized mode.

Functions#

no_grad_eval(f)

Module Contents#

class gigl.src.inference.v1.lib.base_inferencer.BaseInferencer[source]#

Bases: gigl.src.common.types.model.BaseModelOperationsProtocol, Protocol, Generic[T]

The Protocol that you need to implement for your inferencer to function with automated Inference in tabularized mode.

Note: the BaseInferencer class also implements the gigl.src.common.types.model.BaseModelOperationsProtocol protocol, which requires the init_model method, and the getter and setter for the model property.

abstract infer_batch(batch, device=torch.device('cpu'))[source]#

Parameters:

batch (T)
device (torch.device)

Return type:

InferBatchResults

class gigl.src.inference.v1.lib.base_inferencer.InferBatchResults[source]#

embeddings: torch.Tensor | None[source]#

predictions: torch.Tensor | None[source]#

class gigl.src.inference.v1.lib.base_inferencer.NodeAnchorBasedLinkPredictionBaseInferencer[source]#

Bases: BaseInferencer[gigl.src.training.v1.lib.data_loaders.rooted_node_neighborhood_data_loader.RootedNodeNeighborhoodBatch]

The protocol that you need to implement for inderence for Node Anchor Based Link Prediction tasks in tabularized mode. Note: the protocol also implements BaseInferencer protocol.

class gigl.src.inference.v1.lib.base_inferencer.SupervisedNodeClassificationBaseInferencer[source]#

Bases: BaseInferencer[gigl.src.training.v1.lib.data_loaders.supervised_node_classification_data_loader.SupervisedNodeClassificationBatch]

The protocol that you need to implement for inference for Supervised Node Classification tasks in tabularized mode. Note: the protocol also implements BaseInferencer protocol.

gigl.src.inference.v1.lib.base_inferencer.no_grad_eval(f)[source]#

gigl.src.inference.v1.lib.base_inferencer.T[source]#

gigl.src.inference.v1.lib.base_inferencer.logger[source]#