Update README.md

README.md

@@ -15,7 +15,7 @@ Vision transformer baze-sized (ViT base) feature model. Pre-trained with [Masked
 
 ## Training Procedure
 
-ViT-base MAE F1 was pre-trained on the custom dataset containing more than 1 million Formula 1 images from seasons 2021, 2022, 2023 with both racing and not racing scenes. The traing was performed on a cluster of 8 A100 80GB GPUs provided by [Nebius](https://nebius.com/).
+F1 ViT-base MAE was pre-trained on the custom dataset containing more than 1 million Formula 1 images from seasons 2021, 2022, 2023 with both racing and not racing scenes. The traing was performed on a cluster of 8 A100 80GB GPUs provided by [Nebius](https://nebius.com/).
 
 ### Training Hyperparameters
 
@@ -28,6 +28,10 @@ ViT-base MAE F1 was pre-trained on the custom dataset containing more than 1 mil
 
 ## Comparison with ViT-base MAE pre-trained on ImageNet-1K
 
+Comparison of F1 ViT-base MAE and original ViT-base MAE pre-trained on ImageNet-1K (from https://huggingface.co/facebook/vit-mae-base) by reconstruction results on images from Formula 1 domain. Top is ViT-base MAE F1 reconstruction output, bottom is original ViT-base MAE. 
+
+<img src="comparison_1.png" alt="drawing" width="800"/>
+
 ## How to use
 
 Usage is the same as in [Transformers library realization of MAE](https://huggingface.co/facebook/vit-mae-base).