DeepImageEmbedder

class lightautoml.image.image.DeepImageEmbedder(device=torch.device, n_jobs=4, random_state=42, is_advprop=True, model_name='efficientnet-b0', weights_path=None, batch_size=128, verbose=True)[source]

Bases: sklearn.base.TransformerMixin

Transformer for image embeddings.

__init__(device=torch.device, n_jobs=4, random_state=42, is_advprop=True, model_name='efficientnet-b0', weights_path=None, batch_size=128, verbose=True)[source]

Pytorch Dataset for EffNetImageEmbedder.

Parameters
  • device (device) – Torch device.

  • n_jobs – Number of threads for dataloader.

  • random_state – Random seed.

  • is_advprop – Use adversarial training.

  • model_name – Name of effnet model.

  • weights_path (Optional[str]) – Path to saved weights.

  • batch_size (int) – Batch size.

  • verbose (bool) – Verbose data processing.

transform(data)

Calculate image embeddings from pathes.

Parameters

data (Sequence[str]) – Sequence of paths.

Return type

ndarray

Returns

Array of embeddings.