ai-forever
commited on
Commit
•
278d6b4
1
Parent(s):
212894d
Update README.md
Browse files
README.md
CHANGED
@@ -22,6 +22,8 @@ Model was trained by [Sber AI](https://github.com/sberbank-ai) and [SberDevices]
|
|
22 |
|
23 |
The primary proposed method is to modify the sparse transformer's attention mask to better control multi-modalities and up to the next level with "hyper-modality". It allows us to calculate the transitions of modalities in both directions, unlike another similar work DALL-E Transformer, which used only one direction, "text to image". The proposed "image to right text" direction is achieved by extension sparse attention mask to the right for auto-repressively text generation with both image and left text condition.
|
24 |
|
|
|
|
|
25 |
# Authors
|
26 |
|
27 |
+ Alex Shonenkov: [Github](https://github.com/shonenkov), [Kaggle GM](https://www.kaggle.com/shonenkov)
|
|
|
22 |
|
23 |
The primary proposed method is to modify the sparse transformer's attention mask to better control multi-modalities and up to the next level with "hyper-modality". It allows us to calculate the transitions of modalities in both directions, unlike another similar work DALL-E Transformer, which used only one direction, "text to image". The proposed "image to right text" direction is achieved by extension sparse attention mask to the right for auto-repressively text generation with both image and left text condition.
|
24 |
|
25 |
+
![rudolph27b_masks.png](https://s3.amazonaws.com/moonup/production/uploads/1663662426135-5f91b1208a61a359f44e1851.png)
|
26 |
+
|
27 |
# Authors
|
28 |
|
29 |
+ Alex Shonenkov: [Github](https://github.com/shonenkov), [Kaggle GM](https://www.kaggle.com/shonenkov)
|