Commit
·
be11381
1
Parent(s):
2d3b997
Update README.md
Browse files
README.md
CHANGED
|
@@ -1 +1,13 @@
|
|
| 1 |
-
# MoVQGAN
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# MoVQGAN
|
| 2 |
+
|
| 3 |
+
MoVQGAN is a new SOTA model in the image reconstruction problem. This model is based on code from the [VQGAN](https://github.com/CompVis/taming-transformers) repository and modifications from the original [MoVQGAN](https://arxiv.org/pdf/2209.09002.pdf) paper.
|
| 4 |
+
|
| 5 |
+
## Models
|
| 6 |
+
The following table compares the 3 versions of the model SBER-MoVQGAN on the Imagenet dataset in terms of FID, SSIM and PSNR metrics.
|
| 7 |
+
A more detailed description of the experiments and a comparison with other models can be found in the [Habr]() post.
|
| 8 |
+
|
| 9 |
+
|Model|Train steps|FID|SSIM|PSNR|
|
| 10 |
+
|:----|:----|:----|:----|:----|
|
| 11 |
+
| f=8, [SBER-MoVQGAN 67M]((https://huggingface.co/ai-forever/MoVQGAN/resolve/main/movqgan_67M.ckpt)) | 2M | 0,96 | 0,7249 | 26,45 |
|
| 12 |
+
| f=8, [SBER-MoVQGAN 102M](https://huggingface.co/ai-forever/MoVQGAN/resolve/main/movqgan_102M.ckpt)| 2360k | 0,78 | 0,7373 | 26,89 |
|
| 13 |
+
| f=8, [SBER-MoVQGAN 270M](https://huggingface.co/ai-forever/MoVQGAN/resolve/main/movqgan_270M.ckpt) | 1330k | **0,69** | **0,7411** | **27,04** |
|