wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-06-23 00:09:32 +08:00
Code Issues Actions 15 Packages Projects Releases Wiki Activity

Use inverted scaling in parameter

Browse Source
This commit is contained in:
Saquib Alam 2023-08-04 05:11:19 +05:30 committed by GitHub
parent 371b1cd7bc
commit 0445901e97
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
comfy_extras/nodes_latent.py
View File

@ -2,7 +2,7 @@ import torch
from nodes import MAX_RESOLUTION from nodes import MAX_RESOLUTION
# diffusers library scale the random noise # diffusers library scale the random noise
default_vae_scaling_factor = 0.18215 default_vae_scaling_factor = 1.0/0.18215
class NoisyLatentImage: class NoisyLatentImage:
def __init__(self, device="cpu"): def __init__(self, device="cpu"):
@ -12,7 +12,7 @@ class NoisyLatentImage:
def INPUT_TYPES(s): def INPUT_TYPES(s):
return {"required": { return {"required": {
"seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}), "seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
"vae_scaling_factor": ("FLOAT", {"default": default_vae_scaling_factor, "min": 0.01, "max": 1.1, "step": 0.01}), "vae_scaling_factor": ("FLOAT", {"default": default_vae_scaling_factor, "min": 0.0, "max": 10, "step": 0.01}),
"width": ("INT", {"default": 512, "min": 64, "max": MAX_RESOLUTION, "step": 8}), "width": ("INT", {"default": 512, "min": 64, "max": MAX_RESOLUTION, "step": 8}),
"height": ("INT", {"default": 512, "min": 64, "max": MAX_RESOLUTION, "step": 8}), "height": ("INT", {"default": 512, "min": 64, "max": MAX_RESOLUTION, "step": 8}),
"batch_size": ("INT", {"default": 1, "min": 1, "max": 64})}} "batch_size": ("INT", {"default": 1, "min": 1, "max": 64})}}
@ -23,7 +23,7 @@ class NoisyLatentImage:
def generate(self, seed, vae_scaling_factor, width, height, batch_size=1): def generate(self, seed, vae_scaling_factor, width, height, batch_size=1):
generator = torch.manual_seed(seed) generator = torch.manual_seed(seed)
latent = torch.randn([batch_size, 4, height // 8, width // 8], generator=generator, device=self.device) / vae_scaling_factor latent = torch.randn([batch_size, 4, height // 8, width // 8], generator=generator, device=self.device) * vae_scaling_factor
return ({"samples":latent}, ) return ({"samples":latent}, )