SAVRN
Search Contact SAVRN

Open-weight model

adetailer

by Dowon Hwang Bingsu/adetailer

coco2017 (only person) Since getattr is classified as a dangerous pickle function, any segmentation model that uses it is classified as unsafe.

Parameters
Context
Weights328.3 MB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads9.8M

SAVRN's Notes on adetailer

Cleanup passes in an image pipeline are where these weights earn their keep. Dowon Hwang publishes them for the ultralytics library, trained on wider_face and skytnt/anime-segmentation, so the job is detecting faces and segmenting anime images, not generating pictures. No runs_on figures exist, and the file size explains why: 14 files, 328 MB. The hardware question is how much headroom remains on the card already running your generation model.

Apache 2.0 keeps the deployment simple: commercial use, modification and redistribution are permitted, provided the license notices and NOTICE file stay intact and significant changes are stated. Two checks before committing. The files ship in pytorch format only, and the publisher's note says the segmentation models get classified as unsafe because getattr counts as a dangerous pickle function, so take them from the publisher and nowhere else. The last update was November 21, 2024, so pin the version you validate.

Model Card

By Dowon Hwang, published under apache-2.0, revision 53cc19de3820.

YOLOv8 Detection Model

Datasets

Face

Hand

Person

deepfashion2

id label
0 short_sleeved_shirt
1 long_sleeved_shirt
2 short_sleeved_outwear
3 long_sleeved_outwear
4 vest
5 sling
6 shorts
7 trousers
8 skirt
9 short_sleeved_dress
10 long_sleeved_dress
11 vest_dress
12 sling_dress

Info

Read the full model card (249 words)

Identity and Version

Repository
Bingsu/adetailer
Publisher
Dowon Hwang
Task
Not stated by the source
Modality
Other
Library
ultralytics
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
53cc19de382014514d9d4038601d261a7faa9b7b
First published
2023-04-26
Last updated
2024-11-21

Files and Weights

14 files, 328.3 MB in total. The weights are 12 files totalling 328.3 MB in pt.

Weights12 files · 328.3 MB
Documentation1 file · 4.0 KB
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
deepfashion2_yolov8s-seg.ptWeights23.9 MB c319c15e8644
face_yolov8m.ptWeights52.0 MB 717923c19b3f
face_yolov8n.ptWeights6.2 MB 70b640f8f60b
face_yolov8n_v2.ptWeights6.2 MB 8f5f2110f83c
face_yolov8s.ptWeights22.5 MB c7237eff2578
face_yolov9c.ptWeights51.6 MB d02fe493c31e
hand_yolov8n.ptWeights6.2 MB 3991202eb69e
hand_yolov8s.ptWeights22.5 MB 70b540063fbc
hand_yolov9c.ptWeights51.6 MB 6f116f686ef5
person_yolov8m-seg.ptWeights54.8 MB c8ab26f51717
person_yolov8n-seg.ptWeights6.8 MB 38fc8aaae97c
person_yolov8s-seg.ptWeights23.9 MB 53c54aec2239
README.mdDocumentation4.0 KB
.gitattributesRepository1.5 KB

License and Download

License
apache-2.0
Access
Open weights, no gate
Download size
328.3 MB
Download from Dowon Hwang

Released by Dowon Hwang through its official repository on Hugging Face. Read the license.

Built From

  • Trained on (disclosed) skytnt/anime-segmentation
  • Trained on (disclosed) wider_face

Memory Requirements

PrecisionWeights in memory
As published328.3 MB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About adetailer

Can I use adetailer commercially?

Yes. adetailer is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.