LibreResNet34-cls

LibreResNet34-cls

ResNet-34 image classifier (224px, ImageNet-1k, 1000 classes), repackaged for LibreYOLO. ~21.8M parameters.

Source

Derived from huggingface/pytorch-image-models (timm), model resnet34.a1_in1k (the "ResNet Strikes Back" A1 recipe, ImageNet-1k). Architecture: He et al. 2015 (https://arxiv.org/abs/1512.03385). Weights by Ross Wightman and the timm contributors. Copyright (c) Ross Wightman. Licensed under the Apache License 2.0.

Modifications

Learned parameters are unchanged from timm. The checkpoint is metadata-wrapped into the LibreYOLO format (model_family / task / nc / names). LibreYOLO's native ResNet (v1.5) mirrors timm/torchvision module names, so inference is bit-identical to timm (max_abs_diff == 0, verified by the parity test). See weights/convert_resnet_weights.py in the LibreYOLO source repository.

Usage

from libreyolo import LibreYOLO

model = LibreYOLO("LibreResNet34-cls.pt")
result = model.predict("image.jpg")[0]
print(result.probs.top1, result.probs.top5)

License

Apache License 2.0. See the LICENSE and NOTICE files in this repository.