metadata
license: apache-2.0
datasets:
- Yegiiii/deities
language:
- en
metrics:
- accuracy
pipeline_tag: image-classification
tags:
- art
- culture
- heritage
- iconography
Model card for Ideityfy-25
A Clip based ViT (Timm model name: vit_base_patch16_clip_224) image classification model. Trained on Deities-25 dataset.
Model Details
- Model Type: Image classification / feature backbone
- Model Stats:
- Total params: 85,966,105
- Trainable params: 166,681
- Image size: 224 x 224
- LoRA rank: 4
- Dataset: deities-25
Eval Stats
- top-1-acc: 81.61%
- top-3-acc: 93.18%
- top-5-acc: 95.89%