MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessment how to use: python3 frame_detection_inference.py --model path/to/model/weights --input path/to/folder/with/videos --output path/to/folder/for/output --step 30