Instructions to use etset/pixai-tagger-v0.9E with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use etset/pixai-tagger-v0.9E with timm:
import timm model = timm.create_model("hf-hub:etset/pixai-tagger-v0.9E", pretrained=True) - Notebooks
- Google Colab
- Kaggle
PixAI Tagger v0.9 (with EVA02-Large Encoder)
Overview
This model is a modified version of deepghs/pixai-tagger-v0.9-onnx. the EVA02-Large encoder layer was ported from SmilingWolf/wd-eva02-large-tagger-v3 and integrated into this model.
Model Details
- Task: Multilabel Image Classification
- Base Tagger / Head:
pixai-labs/pixai-tagger-v0.9(viadeepghs/pixai-tagger-v0.9-onnx) - Encoder Backbone: EVA02-Large (ported from
SmilingWolf/wd-eva02-large-tagger-v3) - Image Size: 448 x 448
- Total Tags: 13,461 (General: 9,741 / Character: 3,720)
Thresholds
- General tags: 0.3
- Character tags: 0.85
Acknowledgements & Credits
This model would not be possible without the excellent work of the following authors. All credits for the architecture, training, and base weights go to them:
PixAI Labs & DeepGHS: pixai-tagger-v0.9
SmilingWolf:wd-eva02-large-tagger-v3
License Apache-2.0 (Inherited from both base models)
- Downloads last month
- -