xinyun/ComfyUI
1
0
Fork 0
mirror of https://git.datalinker.icu/comfyanonymous/ComfyUI synced 2026-08-14 10:03:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix(comfy/utils.py): fix crop pad_to_size bug

Browse Source
This commit is contained in:
MagicBear 2025-07-04 04:08:33 +08:00
parent ddec938cde
commit 26c55e145c
1 changed files with 9 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
comfy/utils.py
View File

@ -873,7 +873,6 @@ def get_tiled_scale_steps(width, height, tile_x, tile_y, overlap):
cols = 1 if width <= tile_x else math.ceil((width - overlap) / (tile_x - overlap)) cols = 1 if width <= tile_x else math.ceil((width - overlap) / (tile_x - overlap))
return rows * cols return rows * cols
@torch.inference_mode() @torch.inference_mode()
def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_amount=4, def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_amount=4,
out_channels=3, output_device="cpu", downscale=False, out_channels=3, output_device="cpu", downscale=False,
@ -889,18 +888,15 @@ def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_am
index_formulas = [index_formulas] * dims index_formulas = [index_formulas] * dims
def pad_to_size(tensor, target_size): def pad_to_size(tensor, target_size):
"""
Pad tensor to target_size (C, H, W) with zeros
"""
c, h, w = tensor.shape[-3:] c, h, w = tensor.shape[-3:]
pad_h = max(target_size[0] - h, 0) pad_h = max(target_size[0] - h, 0)
pad_w = max(target_size[1] - w, 0) pad_w = max(target_size[1] - w, 0)
if pad_h == 0 and pad_w == 0: if pad_h == 0 and pad_w == 0:
return tensor, (0, 0, 0, 0) return tensor, (0, 0, 0, 0) # (left, right, top, bottom)
padding = [0, pad_w, 0, pad_h] # left, right, top, bottom padding = [0, pad_w, 0, pad_h] # [left, right, top, bottom]
padded = torch.nn.functional.pad(tensor, padding, mode='constant', value=0) padded = torch.nn.functional.pad(tensor, padding, mode='constant', value=0)
return padded, (0, 0, pad_h, pad_w) return padded, (0, pad_w, 0, pad_h) # 明确返回 (left, right, top, bottom)
def get_upscale(dim, val): def get_upscale(dim, val):
up = upscale_amount[dim] up = upscale_amount[dim]
@ -949,7 +945,6 @@ def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_am
for b in range(samples.shape[0]): for b in range(samples.shape[0]):
s = samples[b:b + 1] s = samples[b:b + 1]
# handle entire input fitting in a single tile
if all(s.shape[d + 2] <= tile[d] for d in range(dims)): if all(s.shape[d + 2] <= tile[d] for d in range(dims)):
with torch.no_grad(): with torch.no_grad():
output[b:b + 1] = function(s).to(output_device) output[b:b + 1] = function(s).to(output_device)
@ -992,11 +987,13 @@ def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_am
batched_output = function(batched_input).to(output_device) batched_output = function(batched_input).to(output_device)
for idx, upscaled in enumerate(positions_list): for idx, upscaled in enumerate(positions_list):
ps = batched_output[idx:idx + 1] ps = batched_output[idx:idx+1]
left_pad, right_pad, top_pad, bottom_pad = pad_info_list[idx]
pad_t, pad_b, pad_l, pad_r = pad_info_list[idx] if any(x > 0 for x in (left_pad, right_pad, top_pad, bottom_pad)):
if pad_t > 0 or pad_b > 0 or pad_l > 0 or pad_r > 0: ps = ps[...,
ps = ps[..., pad_t:ps.shape[-2] - pad_b, pad_l:ps.shape[-1] - pad_r] top_pad:ps.shape[-2] - bottom_pad,
left_pad:ps.shape[-1] - right_pad]
mask = torch.ones_like(ps) mask = torch.ones_like(ps)
@ -1075,7 +1072,6 @@ def tiled_scale_multidim(samples, function, tile=(64, 64), overlap=8, upscale_am
return output return output
def tiled_scale(samples, function, tile_x=64, tile_y=64, overlap = 8, upscale_amount = 4, out_channels = 3, output_device="cpu", pbar = None): def tiled_scale(samples, function, tile_x=64, tile_y=64, overlap = 8, upscale_amount = 4, out_channels = 3, output_device="cpu", pbar = None):
return tiled_scale_multidim(samples, function, (tile_y, tile_x), overlap=overlap, upscale_amount=upscale_amount, out_channels=out_channels, output_device=output_device, pbar=pbar) return tiled_scale_multidim(samples, function, (tile_y, tile_x), overlap=overlap, upscale_amount=upscale_amount, out_channels=out_channels, output_device=output_device, pbar=pbar)