Update README.md
Browse files
README.md
CHANGED
|
@@ -4,4 +4,7 @@ license: mit
|
|
| 4 |
|
| 5 |
# Text 2 Image
|
| 6 |
this is text 2 timage model based on ViT trained on 1M timage text pairs. backbone is T5 Encoder and Sana audio tokenizer
|
| 7 |
-
for more information: https://github.com/firdavsus/Text2Image
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
# Text 2 Image
|
| 6 |
this is text 2 timage model based on ViT trained on 1M timage text pairs. backbone is T5 Encoder and Sana audio tokenizer
|
| 7 |
+
for more information: https://github.com/firdavsus/Text2Image
|
| 8 |
+
|
| 9 |
+
# VAE-VQ
|
| 10 |
+
this model takes 128x128 image and compressed it to 16x16 codebook with decent quality reconstrcution loss 0.07 (22 epoch 1M images)!
|