github/InvokeAI
1
1
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI.git synced 2026-02-17 23:21:22 -05:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'spezialspezial-patch-9' into development

Browse Source
This commit is contained in:
Lincoln Stein
2022-11-02 18:35:00 -04:00
parent 8164b6b9cf aa247e68be
commit 240e5486c8
21 changed files with 2768 additions and 136 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ldm/generate.py
View File

@@ -896,7 +896,7 @@ class Generate:
assert os.path.exists(image_path), '** "{image_path}" not found. Please enter the name of an existing image file to mask **'
basename,_ = os.path.splitext(os.path.basename(image_path))
if self.txt2mask is None:
self.txt2mask = Txt2Mask(device = self.device)
self.txt2mask = Txt2Mask(device = self.device, refined=True)
segmented = self.txt2mask.segment(image_path,prompt)
trans = segmented.to_transparent()
inverse = segmented.to_transparent(invert=True)

2
ldm/invoke/restoration/outcrop.py
View File

@@ -41,7 +41,7 @@ class Outcrop(object):
height = extended_image.height,
init_img = extended_image,
strength = 0.90,
image_callback = wrapped_callback,
image_callback = wrapped_callback if image_callback else None,
seam_size = opt.seam_size or 96,
seam_blur = opt.seam_blur or 16,
seam_strength = opt.seam_strength or 0.7,

7
ldm/invoke/txt2mask.py
View File

@@ -36,6 +36,7 @@ from torchvision import transforms
CLIP_VERSION = 'ViT-B/16'
CLIPSEG_WEIGHTS = 'src/clipseg/weights/rd64-uni.pth'
CLIPSEG_WEIGHTS_REFINED = 'src/clipseg/weights/rd64-uni-refined.pth'
CLIPSEG_SIZE = 352
class SegmentedGrayscale(object):
@@ -72,14 +73,14 @@ class Txt2Mask(object):
Create new Txt2Mask object. The optional device argument can be one of
'cuda', 'mps' or 'cpu'.
'''
def __init__(self,device='cpu'):
def __init__(self,device='cpu',refined=False):
print('>> Initializing clipseg model for text to mask inference')
self.device = device
self.model = CLIPDensePredT(version=CLIP_VERSION, reduce_dim=64, )
self.model = CLIPDensePredT(version=CLIP_VERSION, reduce_dim=64, complex_trans_conv=refined)
self.model.eval()
# initially we keep everything in cpu to conserve space
self.model.to('cpu')
self.model.load_state_dict(torch.load(CLIPSEG_WEIGHTS, map_location=torch.device('cpu')), strict=False)
self.model.load_state_dict(torch.load(CLIPSEG_WEIGHTS_REFINED if refined else CLIPSEG_WEIGHTS, map_location=torch.device('cpu')), strict=False)
@torch.no_grad()
def segment(self, image, prompt:str) -> SegmentedGrayscale: