AI/diffusers
1
0
Fork 0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-29 07:22:12 +03:00
Code Activity

remove naive up/down sample

Browse Source
This commit is contained in:
patil-suraj
2022-06-30 11:46:01 +02:00
parent c9bd4d4338
commit 81e7144783
2 changed files with 0 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
src/diffusers/models/resnet.py
View File

@@ -957,19 +957,6 @@ def downsample_2d(x, k=None, factor=2, gain=1):
return upfirdn2d(x, torch.tensor(k, device=x.device), down=factor, pad=((p + 1) // 2, p // 2))
def naive_upsample_2d(x, factor=2):
_N, C, H, W = x.shape
x = torch.reshape(x, (-1, C, H, 1, W, 1))
x = x.repeat(1, 1, 1, factor, 1, factor)
return torch.reshape(x, (-1, C, H * factor, W * factor))
def naive_downsample_2d(x, factor=2):
_N, C, H, W = x.shape
x = torch.reshape(x, (-1, C, H // factor, factor, W // factor, factor))
return torch.mean(x, dim=(3, 5))
class NIN(nn.Module):
def __init__(self, in_dim, num_units, init_scale=0.1):
super().__init__()

13
src/diffusers/models/unet_sde_score_estimation.py
View File

@@ -123,19 +123,6 @@ class Conv2d(nn.Module):
return x
def naive_upsample_2d(x, factor=2):
_N, C, H, W = x.shape
x = torch.reshape(x, (-1, C, H, 1, W, 1))
x = x.repeat(1, 1, 1, factor, 1, factor)
return torch.reshape(x, (-1, C, H * factor, W * factor))
def naive_downsample_2d(x, factor=2):
_N, C, H, W = x.shape
x = torch.reshape(x, (-1, C, H // factor, factor, W // factor, factor))
return torch.mean(x, dim=(3, 5))
def upsample_conv_2d(x, w, k=None, factor=2, gain=1):
"""Fused `upsample_2d()` followed by `tf.nn.conv2d()`.