patrickramos
commited on
Commit
•
7c68a99
1
Parent(s):
70dfc27
Update app.py
Browse files
app.py
CHANGED
@@ -150,7 +150,22 @@ article = '''An extension of [Emb-GAMs](https://arxiv.org/abs/2209.11799), visua
|
|
150 |
|
151 |
The use of a sum of embeddings allows us to visualize which regions of an image contributed positive or negatively to each class score.
|
152 |
|
153 |
-
No paper yet, but you can refer to
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
154 |
|
155 |
demo = gr.Interface(
|
156 |
fn=visualize,
|
|
|
150 |
|
151 |
The use of a sum of embeddings allows us to visualize which regions of an image contributed positive or negatively to each class score.
|
152 |
|
153 |
+
No paper yet, but you can refer to these tweets:
|
154 |
+
|
155 |
+
- [Tweet #1](https://twitter.com/patrick_j_ramos/status/1586992857969147904?s=20&t=5-j5gKK0FpZOgzR_9Wdm1g)
|
156 |
+
- [Tweet #2](https://twitter.com/patrick_j_ramos/status/1602187142062804992?s=20&t=roTFXfMkHHYVoCuNyN-AUA)
|
157 |
+
|
158 |
+
Also, check out the original [Emb-GAM paper](https://arxiv.org/abs/2209.11799).
|
159 |
+
|
160 |
+
```bibtex
|
161 |
+
@article{singh2022emb,
|
162 |
+
title={Emb-GAM: an Interpretable and Efficient Predictor using Pre-trained Language Models},
|
163 |
+
author={Singh, Chandan and Gao, Jianfeng},
|
164 |
+
journal={arXiv preprint arXiv:2209.11799},
|
165 |
+
year={2022}
|
166 |
+
}
|
167 |
+
```
|
168 |
+
'''
|
169 |
|
170 |
demo = gr.Interface(
|
171 |
fn=visualize,
|