Upload folder using huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,3 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# Tibetan Script Classifier (DINOv3)
|
| 2 |
|
| 3 |
This repository contains fine-tuned Tibetan script classification checkpoints for 18 classes, trained from the DINOv3 ViT-S backbone:
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- bo
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
library_name: transformers
|
| 6 |
+
tags:
|
| 7 |
+
- image-classification
|
| 8 |
+
- dinov3
|
| 9 |
+
- tibetan
|
| 10 |
+
- script-classification
|
| 11 |
+
- paleography
|
| 12 |
+
- fine-tuned
|
| 13 |
+
- document-analysis
|
| 14 |
+
base_model: facebook/dinov3-vits16-pretrain-lvd1689m
|
| 15 |
+
datasets:
|
| 16 |
+
- custom
|
| 17 |
+
metrics:
|
| 18 |
+
- f1
|
| 19 |
+
- accuracy
|
| 20 |
+
pipeline_tag: image-classification
|
| 21 |
+
model-index:
|
| 22 |
+
- name: Tibetan Script Classifier (DINOv3 ViT-S)
|
| 23 |
+
results:
|
| 24 |
+
- task:
|
| 25 |
+
type: image-classification
|
| 26 |
+
name: Tibetan Script Classification
|
| 27 |
+
metrics:
|
| 28 |
+
- name: Macro F1 (whole page)
|
| 29 |
+
type: f1
|
| 30 |
+
value: 0.512
|
| 31 |
+
- name: Accuracy (whole page)
|
| 32 |
+
type: accuracy
|
| 33 |
+
value: 0.571
|
| 34 |
+
- name: Macro F1 (CLAHE patches, page-level)
|
| 35 |
+
type: f1
|
| 36 |
+
value: 0.529
|
| 37 |
+
---
|
| 38 |
+
|
| 39 |
# Tibetan Script Classifier (DINOv3)
|
| 40 |
|
| 41 |
This repository contains fine-tuned Tibetan script classification checkpoints for 18 classes, trained from the DINOv3 ViT-S backbone:
|