avans06's picture
The inference uses the auto_split_upscale mechanism.
c85d0ce
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# The file source is from the [ESRGAN](https://github.com/xinntao/ESRGAN) project
# forked by authors [joeyballentine](https://github.com/joeyballentine/ESRGAN) and [BlueAmulet](https://github.com/BlueAmulet/ESRGAN).
import gc
import numpy as np
import torch
def bgr_to_rgb(image: torch.Tensor) -> torch.Tensor:
# flip image channels
# https://github.com/pytorch/pytorch/issues/229
out: torch.Tensor = image.flip(-3)
# out: torch.Tensor = image[[2, 1, 0], :, :] #RGB to BGR #may be faster
return out
def rgb_to_bgr(image: torch.Tensor) -> torch.Tensor:
# same operation as bgr_to_rgb(), flip image channels
return bgr_to_rgb(image)
def bgra_to_rgba(image: torch.Tensor) -> torch.Tensor:
out: torch.Tensor = image[[2, 1, 0, 3], :, :]
return out
def rgba_to_bgra(image: torch.Tensor) -> torch.Tensor:
# same operation as bgra_to_rgba(), flip image channels
return bgra_to_rgba(image)
def auto_split_upscale(
lr_img: np.ndarray,
upscale_function,
scale: int = 4,
overlap: int = 32,
max_depth: int = None,
current_depth: int = 1,
):
# Attempt to upscale if unknown depth or if reached known max depth
if max_depth is None or max_depth == current_depth:
try:
print(f"auto_split_upscale, current depth: {current_depth}")
result, _ = upscale_function(lr_img, scale)
return result, current_depth
except RuntimeError as e:
# Check to see if its actually the CUDA out of memory error
if "CUDA" in str(e):
# Collect garbage (clear VRAM)
torch.cuda.empty_cache()
gc.collect()
# Re-raise the exception if not an OOM error
else:
raise RuntimeError(e)
finally:
# Free GPU memory and clean up resources
torch.cuda.empty_cache()
gc.collect()
h, w, c = lr_img.shape
# Split image into 4ths
top_left = lr_img[: h // 2 + overlap, : w // 2 + overlap, :]
top_right = lr_img[: h // 2 + overlap, w // 2 - overlap :, :]
bottom_left = lr_img[h // 2 - overlap :, : w // 2 + overlap, :]
bottom_right = lr_img[h // 2 - overlap :, w // 2 - overlap :, :]
# Recursively upscale the quadrants
# After we go through the top left quadrant, we know the maximum depth and no longer need to test for out-of-memory
top_left_rlt, depth = auto_split_upscale(
top_left,
upscale_function,
scale=scale,
overlap=overlap,
max_depth=max_depth,
current_depth=current_depth + 1,
)
top_right_rlt, _ = auto_split_upscale(
top_right,
upscale_function,
scale=scale,
overlap=overlap,
max_depth=depth,
current_depth=current_depth + 1,
)
bottom_left_rlt, _ = auto_split_upscale(
bottom_left,
upscale_function,
scale=scale,
overlap=overlap,
max_depth=depth,
current_depth=current_depth + 1,
)
bottom_right_rlt, _ = auto_split_upscale(
bottom_right,
upscale_function,
scale=scale,
overlap=overlap,
max_depth=depth,
current_depth=current_depth + 1,
)
# Define output shape
out_h = h * scale
out_w = w * scale
# Create blank output image
output_img = np.zeros((out_h, out_w, c), np.uint8)
# Fill output image with tiles, cropping out the overlaps
output_img[: out_h // 2, : out_w // 2, :] = top_left_rlt[
: out_h // 2, : out_w // 2, :
]
output_img[: out_h // 2, -out_w // 2 :, :] = top_right_rlt[
: out_h // 2, -out_w // 2 :, :
]
output_img[-out_h // 2 :, : out_w // 2, :] = bottom_left_rlt[
-out_h // 2 :, : out_w // 2, :
]
output_img[-out_h // 2 :, -out_w // 2 :, :] = bottom_right_rlt[
-out_h // 2 :, -out_w // 2 :, :
]
return output_img, depth