You can deploy this model and start earning money today!

aaristov/
bacterial-morphology-resnet34

Name: bacterial-morphology-resnet34
Availability: InStock

Views

No views yet

Bacterial Morphology ResNet34

Multi-label classifier of bacterial droplet morphology from the paper Extending digital biology: bacterial survival and morphological heterogeneity under antibiotic stress (Maikranz et al.). Source data & weights: https://doi.org/10.5281/zenodo.17193389 · Code: https://github.com/BaroudLab/Bacterial-Morphology

Backbone: torchvision resnet34, final layer nn.Linear(512, 5)
Weights file: bacteria_trained_model_resnet34 (plain state_dict)
Classes (order): positive, planktonic, clumped, rods, filaments
Multi-label: trained with BCEWithLogitsLoss -> apply sigmoid, threshold 0.5
Preprocessing: Grayscale(3) -> Resize(224) -> ToTensor -> Normalize(mean=[0.485,0.456,0.406], std=[0.229,0.224,0.225])

Load

python

import torch, torchvision
m = torchvision.models.resnet34()
m.fc = torch.nn.Linear(512, 5)
m.load_state_dict(torch.load("bacteria_trained_model_resnet34", map_location="cpu"))
m.eval()

aaristov/bacterial-morphology-resnet34

Bacterial Morphology ResNet34

Load

aaristov/
bacterial-morphology-resnet34