meerqat.models.image module#

Building blocks for computer vision models.

class meerqat.models.image.FaceEmbedding(face_dim, bbox_dim, embedding_dim, dropout=0.1, layer_norm_eps=1e-12)[source]#

Bases: Module

Projects a face feature in the embedding space using a linear layer together with the corresponding bounding box.

forward(face, bbox, image_type_embeddings=None)[source]#

Defines 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.

class meerqat.models.image.ImageEmbedding(input_dim, embedding_dim, dropout=0.1, layer_norm_eps=None)[source]#

Bases: Module

Projects an image feature in the embedding space using a linear layer.

forward(input, image_type_embeddings=None)[source]#

Defines 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.