aesthetic-predictor - GitLite

KO | EN

gitlite — search

#typescript #kotlin #nextjs #hacktoberfest #computer-vision #large-language-models #python #markdown #deep-learning #css

aesthetic-predictor

(★ 730)

A linear estimator on top of clip to predict the aesthetic quality of pictures

File Explorer

ZIP Download

- rating0.npy
- rating1.npy
- rating2.npy
- rating3.npy
- rating4.npy
- rating5.npy
- rating6.npy
- rating7.npy
- rating8.npy
- rating9.npy
- rating0.npy
- rating1.npy
- rating2.npy
- rating3.npy
- rating4.npy
- rating5.npy
- rating6.npy
- rating7.npy
- rating8.npy
- rating9.npy
asthetics_predictor.ipynb
cat_aesthetic.png
cat_normal.png
LICENSE
README.md
sa_0_4_vit_b_16_linear.pth
sa_0_4_vit_b_32_linear.pth
sa_0_4_vit_l_14_linear.pth

# Use via CDN

jsDelivr

jsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.

Version

File

Link

Example

# Project Badges

Open In Colab Visit

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

A linear estimator on top of clip to predict the aesthetic quality of pictures

Normal results:

<img src="cat_normal.png" width="512" />

Aesthetic results:

<img src="cat_aesthetic.png" width="512" />

visualization over laion2B http://3080.rom1504.fr/aesthetic/aesthetic_viz.html

integrated into https://rom1504.github.io/clip-retrieval/

```python
import os
import torch
import torch.nn as nn
from os.path import expanduser  # pylint: disable=import-outside-toplevel
from urllib.request import urlretrieve  # pylint: disable=import-outside-toplevel
def get_aesthetic_model(clip_model="vit_l_14"):
    """load the aethetic model"""
    home = expanduser("~")
    cache_folder = home + "/.cache/emb_reader"
    path_to_model = cache_folder + "/sa_0_4_"+clip_model+"_linear.pth"
    if not os.path.exists(path_to_model):
        os.makedirs(cache_folder, exist_ok=True)
        url_model = (
            "https://github.com/LAION-AI/aesthetic-predictor/blob/main/sa_0_4_"+clip_model+"_linear.pth?raw=true"
        )
        urlretrieve(url_model, path_to_model)
    if clip_model == "vit_l_14":
        m = nn.Linear(768, 1)
    elif clip_model == "vit_b_32":
        m = nn.Linear(512, 1)
    else:
        raise ValueError()
    s = torch.load(path_to_model)
    m.load_state_dict(s)
    m.eval()
    return m
```

That model takes clip embeddings as input

https://github.com/rom1504/embedding-reader/blob/main/examples/aesthetic_inference.py
https://github.com/rom1504/embedding-reader/blob/main/examples/aesthetic_inference_emb.py