EventStream.transformer.lightning_modules.embedding module¶

A PyTorch Lightning runnable module for getting embeddings from a pre-trained ESGPT model.

class EventStream.transformer.lightning_modules.embedding.ESTForEmbedding(config: StructuredTransformerConfig | dict[str, Any], pretrained_weights_fp: Path)[source]¶

Bases: LightningModule

A PyTorch Lightning Module for extracting embeddings only model.

predict_step(batch, batch_idx)[source]¶

Retrieves the embeddings and returns them.

class EventStream.transformer.lightning_modules.embedding.EmbeddingsOnlyModel(config: StructuredTransformerConfig)[source]¶

Bases: StructuredTransformerPreTrainedModel

A “model” which simply retrieves the final embeddings from the encoder and returns them.

forward(*args, **kwargs)[source]¶

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

EventStream.transformer.lightning_modules.embedding.get_embeddings(cfg: FinetuneConfig)[source]¶

Gets embeddings.

Writes embeddings to cfg.load_from_model_dir / "embeddings" / cfg.task_df_name / "{split}_embeddings.pt".

Parameters:¶

cfg: FinetuneConfig¶

The fine-tuning configuration object specifying the cohort for which and model from which you wish to get embeddings.