Skip to content

perf(nodes): avoid two full-size copies in Spandrel upscale #9426

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
dexhunter wants to merge 1 commit into
base: main
Choose a base branch
from
Open

perf(nodes): avoid two full-size copies in Spandrel upscale #9426

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 12 additions & 12 deletions invokeai/app/invocations/spandrel_image_to_image.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,6 @@
import functools
from typing import Callable

import numpy as np
import torch
from PIL import Image
from tqdm import tqdm
Expand Down Expand Up @@ -90,20 +89,17 @@ def upscale_image(
tiles = sorted(tiles, key=lambda x: x.coords.left)
tiles = sorted(tiles, key=lambda x: x.coords.top)

# Prepare input image for inference.
image_tensor = SpandrelImageToImageModel.pil_to_tensor(image)

# Scale the tiles for re-assembling the final image.
scale = spandrel_model.scale
scaled_tiles = [cls.scale_tile(tile, scale=scale) for tile in tiles]

# Prepare the output tensor.
_, channels, height, width = image_tensor.shape
channels = len(image.getbands())
output_tensor = torch.zeros(
(height * scale, width * scale, channels), dtype=torch.uint8, device=torch.device("cpu")
(image.height * scale, image.width * scale, channels), dtype=torch.uint8, device=torch.device("cpu")
)

image_tensor = image_tensor.to(device=TorchDevice.choose_torch_device(), dtype=spandrel_model.dtype)
device = TorchDevice.choose_torch_device()

# Run the model on each tile.
pbar = tqdm(list(zip(tiles, scaled_tiles, strict=True)), desc="Upscaling Tiles")
Expand All @@ -116,8 +112,12 @@ def upscale_image(
if is_canceled():
raise CanceledException

# Extract the current tile from the input tensor.
input_tile = image_tensor[:, :, tile.coords.top : tile.coords.bottom, tile.coords.left : tile.coords.right]
# Crop the current tile from the input image and convert it on demand. Converting the
# whole image up front would keep a float32 copy of it in memory for the entire loop,
# even though only one tile is ever used at a time.
input_tile = SpandrelImageToImageModel.pil_to_tensor(
image.crop((tile.coords.left, tile.coords.top, tile.coords.right, tile.coords.bottom))
).to(device=device, dtype=spandrel_model.dtype)

# Run the model on the tile.
output_tile = spandrel_model.run(input_tile)
Expand All @@ -144,9 +144,9 @@ def upscale_image(

step_callback(pbar.n + 1, pbar.total)

# Convert the output tensor to a PIL image.
np_image = output_tensor.detach().numpy().astype(np.uint8)
pil_image = Image.fromarray(np_image)
# Convert the output tensor to a PIL image. `output_tensor` is already uint8, so `.numpy()`
# is a zero-copy view; casting it to uint8 again here would copy the whole image for nothing.
pil_image = Image.fromarray(output_tensor.numpy())

return pil_image

Expand Down
Loading