pyiqa.archs.metaiqa_arch¶

MetaIQA Model wrapper for pyiqa. Reference: Zhu, Hancheng, et al. “MetaIQA: Deep Meta-Learning for No-Reference Image Quality Assessment.” CVPR 2020. Re-implemented and optimized for pyiqa ecosystem.

Module Contents¶

pyiqa.archs.metaiqa_arch.default_model_urls[source]¶

class pyiqa.archs.metaiqa_arch.BaselineModel1(num_classes=1, keep_probability=0.5, inputsize=1000)[source]¶

Bases: torch.nn.Module

forward(x)[source]¶

class pyiqa.archs.metaiqa_arch.MetaIQA(pretrained=True, pretrained_model_path=None, **kwargs)[source]¶

Bases: torch.nn.Module

preprocess(x)[source]¶

forward(x, ref=None)[source]¶