v1.1: held-out region and background models

Four ResNet-50 models (448 px) trained on the held-out cluster splits of Experiments 3 and 4. Same recipe as v1.0: SGD (Nesterov, momentum 0.9, lr 0.01, head lr x10, weight decay 1e-4), cosine schedule, 30 epochs, torchvision ImageNet weights (resnet50.tv_in1k), float16 mixed precision. Best checkpoint by val accuracy; weights only.

File	Split	Mode	Epoch	Val	Held-out test	SHA-256
resnet50_448_geography_full.pth	Geography (iNaturalist, 156 species)	full image	28	81.94%	62.68% (western N. America, 18,953)	527404323a41a8cc331489c29dedaacd483adcb102b58cee58c67f113ee9e703
resnet50_448_bgsplit_full.pth	Background clusters (CUB)	full image	25	85.66%	83.96% (ground / grass, 2,350)	9501f27a4e4a7927c7fdf5d5de76e61404a03752564a1011717f136719b43dd7
resnet50_448_bgsplit_crop.pth	Background clusters (CUB)	bird crop	23	86.93%	85.49%	d0449c02294d400aca99b1467b3aa9fb0348a793fbb85545c0b2a2f473ab49b2
resnet50_448_bgsplit_birdshift.pth	Background clusters (CUB)	BirdShift	30	86.65%	84.30%	810a41f863992b8a70016ee9ec34df878d3d63a111f8c82c7bf81855ec202228

These models are trained on the split manifests in data/splits/, not on CUB's official train/test split, so they should not be scored on the official CUB test set (the background split puts about half of the official test images into training). The best model on the official split is still birdshift_resnet50_448_cub.pth in v1.0 (full-image, 84.55%).

Load: BirdResNet(num_classes=200, pretrained=False).load_state_dict(ck["model_state_dict"]). Each file also stores mode, image_size and manifest. Evaluate with scripts/eval_test.py --manifest <split csv> --splits val test (copy the file to checkpoints/<run>/best.pth). Full-image and BirdShift take Resize(512) + CenterCrop(448); the crop model takes the ground-truth box crop, square-padded and resized to 448.

v1.0: ResNet-50, 84.55% on CUB-200-2011 test

Full-image ResNet-50 at 448 px, fine-tuned on all 5,994 CUB-200-2011 training images.

Top-1 on the official test split: 84.55% (4,899 / 5,794).

Recipe: SGD (Nesterov, momentum 0.9, lr 0.01, head lr x10, weight decay 1e-4), cosine schedule, 50 epochs, torchvision ImageNet weights (resnet50.tv_in1k), Resize(512) + RandomCrop(448) + horizontal flip, float16 mixed precision.

File: birdshift_resnet50_448_cub.pth (weights only)
SHA-256: a647b1bbb444ac05985a3bbc2443e79bfb437ee7e83140e1fdafaa4788c1869c

See the README section "Matching the published ResNet-50 number" for loading code and the full comparison.

Note: this is the full-image model; "birdshift" in the file name is the project name. Held-out split models (including BirdShift) are in v1.1.