superduper-torch 0.4.0

Superduper allows users to work with arbitrary `torch` models, with custom pre-, post-processing and input/ output data-types, as well as offering training with superduper

superduper_torch

Superduper allows users to work with arbitrary torch models, with custom pre-, post-processing and input/ output data-types, as well as offering training with superduper

Installation

pip install superduper_torch

API

• Code
• API-docs

Class	Description
superduper_torch.model.TorchModel	Torch model. This class is a wrapper around a PyTorch model.
superduper_torch.training.TorchTrainer	Configuration for the PyTorch trainer.

Examples

TorchModel

import torch
from superduper_torch.model import TorchModel

model = TorchModel(
    object=torch.nn.Linear(32, 1),
    identifier="test",
    preferred_devices=("cpu",),
    postprocess=lambda x: int(torch.sigmoid(x).item() > 0.5),
)
model.predict(torch.randn(32))

Training Example

Read more about this here