Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
license_name: sam-license
|
| 4 |
+
license_link: LICENSE
|
| 5 |
+
base_model: facebook/sam3
|
| 6 |
+
---
|
| 7 |
+
|
| 8 |
+
# SlideForge SAM3 slide-component decoder
|
| 9 |
+
|
| 10 |
+
Fine-tuned SAM3 decoder for perception-aligned slide decomposition
|
| 11 |
+
(30.4M trainable parameters, 3.6% of the model; frozen vision-language
|
| 12 |
+
backbone). Trained on 127 academic decks / 3,510 slides / 25,132
|
| 13 |
+
component boxes over a 306-class open-vocabulary slide-component
|
| 14 |
+
taxonomy. Mean IoU 0.873 on a slide-disjoint held-out split.
|
| 15 |
+
|
| 16 |
+
Use with the base `facebook/sam3` checkpoint via the SlideForge
|
| 17 |
+
pipeline: https://github.com/UIUC-MONET/SLIDEFORGE
|
| 18 |
+
|
| 19 |
+
Paper: SlideForge: An LLM Agent for Controllable Editing of Slides as
|
| 20 |
+
Structured Artifacts.
|