---
title: wd-v1-4-moat-tagger-v2
canonical_url: "https://www.modelscope.cn/models/fireicewolf/wd-v1-4-moat-tagger-v2"
md_url: "https://www.modelscope.cn/models/fireicewolf/wd-v1-4-moat-tagger-v2.md"
repository: fireicewolf/wd-v1-4-moat-tagger-v2
last_updated: 2024-04-20
license: apache-2.0
pipeline_tag: image-classification
tasks:
  - image-classification
library_name:
  - onnx
  - other
frameworks:
  - other
downloads: 634
stars: 1
---

# wd-v1-4-moat-tagger-v2

> wd-v1-4-moat-tagger-v2 - fireicewolf 在 ModelScope 开源的模型。This is a fork version from https://huggingface.co/SmilingWolf/wd-v1-4-moat-tagger-v2

fireicewolf/wd-v1-4-moat-tagger-v2 是 ModelScope 魔搭社区上的image-classification模型，采用 apache-2.0 许可。

- **Repository**: fireicewolf/wd-v1-4-moat-tagger-v2
- **License**: apache-2.0
- **Tasks**: image-classification
- **Downloads**: 634
- **Stars**: 1
- **Last updated**: 2024-04-20

Source: https://www.modelscope.cn/models/fireicewolf/wd-v1-4-moat-tagger-v2

---

# WD 1.4 MOAT Tagger V2

This is a fork version from https://huggingface.co/SmilingWolf/wd-v1-4-moat-tagger-v2

Supports ratings, characters and general tags.

Trained using https://github.com/SmilingWolf/SW-CV-ModelZoo.  
TPUs used for training kindly provided by the [TRC program](https://sites.research.google/trc/about/).

## Dataset
Last image id: 5944504  
Trained on Danbooru images with IDs modulo 0000-0899.  
Validated on images with IDs modulo 0950-0999.  
Images with less than 10 general tags were filtered out.  
Tags with less than 600 images were filtered out.

## Validation results
`P=R: threshold = 0.3771, F1 = 0.6911`

## Paper
[`MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models`](https://arxiv.org/abs/2210.01820)

## Final words
Subject to change and updates.  
Downstream users are encouraged to use tagged releases rather than relying on the head of the repo.