plismbench.models.mahmood_lab module

Models from Mahmood Lab.

class plismbench.models.mahmood_lab.UNI(device: int | list[int] | None = -1, mixed_precision: bool = False)

Bases: Extractor

UNI model developped by Mahmood Lab available on Hugging-Face (1).

Note

1. https://huggingface.co/MahmoodLab/UNI

Parameters:

• device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

• mixed_precision (bool = True) – Whether to use mixed_precision.

property transform: Compose

Transform method to apply element wise.

class plismbench.models.mahmood_lab.UNI2h(device: int | list[int] | None = -1, mixed_precision: bool = False)

Bases: Extractor

UNI2-h model developped by Mahmood Lab available on Hugging-Face (1).

Note

1. https://huggingface.co/MahmoodLab/UNI2-h

Parameters:

• device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

• mixed_precision (bool = True) – Whether to use mixed_precision.

property transform: Compose

Transform method to apply element wise.

class plismbench.models.mahmood_lab.CONCH(device: int | list[int] | None = -1, mixed_precision: bool = False)

Bases: Extractor

CONCH model developped by Mahmood Lab available on Hugging-Face (1).

Note

1. https://huggingface.co/MahmoodLab/CONCH

Parameters:

• device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

• mixed_precision (bool = True) – Whether to use mixed_precision.

process(image) → Tensor

Process input images.

property transform: Lambda

Transform method to apply element wise.

class plismbench.models.mahmood_lab.CONCHv15(device: int | list[int] | None = -1, mixed_precision: bool = False)

Bases: Extractor

Conchv15 model available from TITAN on Hugging-Face (1).

Note

1. https://huggingface.co/MahmoodLab/conchv1_5

property transform: Lambda

Transform method to apply element wise.