Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 

hat

Input

Input

(Image from https://github.com/JingyunLiang/SwinIR/tree/main/testsets)

Ailia input shape : (1, 3, 256, 256)

Output

Output

Ailia output shape : (1, 3, 256 * scale, 256 * scale)

default : scale=2

Usage

Automatically downloads the onnx and prototxt files when running. It is necessary to be connected to the Internet while downloading.

For the sample image with twice the resolution (BI),

$ python3 hat.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to chatge the name of the output file to save.

$ python3 hat.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

["HAT","HAT_S","HAT_GAN_REAL_sharper","HAT_GAN_REAL"],

By adding the --arch option, you can specify model type which is selected from "HAT","HAT_S","HAT_GAN_REAL_sharper","HAT_GAN_REAL".
(default is HAT)

$ python3 hat.py --arch HAT

If you want to specify the scale for the resolution, put the scale after the --scale option.
Choose the scale in [2, 3, 4].

$ python3 hat.py --scale SCALE 

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 hat.py --video VIDEO_PATH

arch 説明

Reference

Hat

Framework

Pytorch 1.10.0

Model Format

ONNX opset = 11

Netron

Hat_S_x2.onnx.prototxt

Hat_S_x3.onnx.prototxt

Hat_S_x4.onnx.prototxt

Hat_x2.onnx.prototxt

Hat_x3.onnx.prototxt

Hat_x4.onnx.prototxt

HAT_GAN_REAL.onnx.prototxt

HAT_GAN_REAL_sharper.onnx.prototxt