ANGExllL/obsidian

EfficientNetV2-L (torchvision efficientnet_v2_l, 1000 ImageNet classes) fine-tuned with adversarial training for the Perturb subnet (netuid 26).

• miner hotkey: 5Ebm5cRFAzkHGtq3myyvnL3w6GVSPrAy4LY6HSFMNdQLnomj
• on-chain hash: sha256(model.safetensors || hotkey) = 04ca6700a69524b39e51648d5db5e44947026844fd7d6b109c34b1b9326da00f
• preprocessing: EfficientNet_V2_L_Weights.IMAGENET1K_V1.transforms() (bicubic resize 480, center crop 480, mean = std = 0.5)

from safetensors.torch import load_file
from torchvision.models import efficientnet_v2_l
model = efficientnet_v2_l(weights=None)
model.load_state_dict(load_file("model.safetensors"))