search

IceClear / CLIP-IQA

[AAAI 2023] Exploring CLIP for Assessing the Look and Feel of Images
Other
342 stars 20 forks source link

Model loading #9

Open yongliuy opened 1 year ago

yongliuy commented 1 year ago

When we try to run clipiqa_single_image_demo.py and load pre-trained model, there will be the following problems:

unexpected key in source state_dict: generator.clipmodel_0.logit_scale, generator.clipmodel_0.prompt_learner.ctx,....... missing keys in source state_dict: generator.clip_model.positional_embedding, generator.clip_model.text_projection,......

How should I solve it?

IceClear commented 1 year ago

5