Add pipeline tag, library name, and project page link
Browse filesThis PR adds the missing `pipeline_tag` and `library_name` to the model card metadata. The `pipeline_tag` is set to `image-text-to-text` given the model's functionality. The `library_name` is set to `open_clip` based on the provided code examples and file structure.
README.md
CHANGED
@@ -1,11 +1,13 @@
|
|
1 |
---
|
2 |
-
license: cc-by-4.0
|
3 |
datasets:
|
4 |
- UCSC-VLAA/Recap-DataComp-1B
|
5 |
- mlfoundations/datacomp_1b
|
6 |
library_name: open_clip
|
|
|
|
|
7 |
---
|
8 |
-
|
|
|
9 |
|
10 |
A DeltaCLIP-H/14-336 Model that is adversarially pre-trained with web-scale image-text data to reach non-robust-VLM helpfulness levels on clean data while being robust on adversarially attacked data.
|
11 |
|
|
|
1 |
---
|
|
|
2 |
datasets:
|
3 |
- UCSC-VLAA/Recap-DataComp-1B
|
4 |
- mlfoundations/datacomp_1b
|
5 |
library_name: open_clip
|
6 |
+
license: cc-by-4.0
|
7 |
+
pipeline_tag: image-text-to-text
|
8 |
---
|
9 |
+
|
10 |
+
[[Paper]](https://arxiv.org/abs/2501.09446) [[github]](https://github.com/zw615/Double_Visual_Defense) [[Project Page]](https://doublevisualdefense.github.io/)
|
11 |
|
12 |
A DeltaCLIP-H/14-336 Model that is adversarially pre-trained with web-scale image-text data to reach non-robust-VLM helpfulness levels on clean data while being robust on adversarially attacked data.
|
13 |
|