In the config.yaml file, your model seems like are trained on 256x256 image size.
But the resolution in paper is 1280x720.

Can you provided the config.yaml which is used to train the best model on GROPRO dataset?

Another question is about the image augmentations

Is the image augmentations provided in the config.yaml file the best practice?

"""
size: &SIZE 256
crop: random
preload: &PRELOAD false
preload_size: &PRELOAD_SIZE 0
bounds: [0, .9]
scope: geometric
corrupt: &CORRUPT
- name: cutout
prob: 0.5
num_holes: 3
max_h_size: 25
max_w_size: 25
- name: jpeg
quality_lower: 70
quality_upper: 90
- name: motion_blur
- name: median_blur
- name: gamma
- name: rgb_shift
- name: hsv_shift
- name: sharpen
"""