add --no-half arg
This commit is contained in:
parent
6cfc7c30f1
commit
59ee89bd34
@ -74,7 +74,6 @@ const Toast = React.forwardRef<
|
|||||||
})
|
})
|
||||||
|
|
||||||
Toast.defaultProps = {
|
Toast.defaultProps = {
|
||||||
desc: '',
|
|
||||||
state: 'loading',
|
state: 'loading',
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -15,8 +15,9 @@ class PaintByExample(InpaintModel):
|
|||||||
min_size = 512
|
min_size = 512
|
||||||
|
|
||||||
def init_model(self, device: torch.device, **kwargs):
|
def init_model(self, device: torch.device, **kwargs):
|
||||||
|
fp16 = not kwargs['no_half']
|
||||||
use_gpu = device == torch.device('cuda') and torch.cuda.is_available()
|
use_gpu = device == torch.device('cuda') and torch.cuda.is_available()
|
||||||
torch_dtype = torch.float16 if use_gpu else torch.float32
|
torch_dtype = torch.float16 if use_gpu and fp16 else torch.float32
|
||||||
self.model = DiffusionPipeline.from_pretrained(
|
self.model = DiffusionPipeline.from_pretrained(
|
||||||
"Fantasy-Studio/Paint-by-Example",
|
"Fantasy-Studio/Paint-by-Example",
|
||||||
torch_dtype=torch_dtype,
|
torch_dtype=torch_dtype,
|
||||||
|
@ -30,6 +30,7 @@ class SD(InpaintModel):
|
|||||||
|
|
||||||
def init_model(self, device: torch.device, **kwargs):
|
def init_model(self, device: torch.device, **kwargs):
|
||||||
from diffusers.pipelines.stable_diffusion import StableDiffusionInpaintPipeline
|
from diffusers.pipelines.stable_diffusion import StableDiffusionInpaintPipeline
|
||||||
|
fp16 = not kwargs['no_half']
|
||||||
|
|
||||||
model_kwargs = {"local_files_only": kwargs['sd_run_local']}
|
model_kwargs = {"local_files_only": kwargs['sd_run_local']}
|
||||||
if kwargs['sd_disable_nsfw']:
|
if kwargs['sd_disable_nsfw']:
|
||||||
@ -42,7 +43,7 @@ class SD(InpaintModel):
|
|||||||
torch_dtype = torch.float16 if use_gpu else torch.float32
|
torch_dtype = torch.float16 if use_gpu else torch.float32
|
||||||
self.model = StableDiffusionInpaintPipeline.from_pretrained(
|
self.model = StableDiffusionInpaintPipeline.from_pretrained(
|
||||||
self.model_id_or_path,
|
self.model_id_or_path,
|
||||||
revision="fp16" if use_gpu else "main",
|
revision="fp16" if use_gpu and fp16 else "main",
|
||||||
torch_dtype=torch_dtype,
|
torch_dtype=torch_dtype,
|
||||||
use_auth_token=kwargs["hf_access_token"],
|
use_auth_token=kwargs["hf_access_token"],
|
||||||
**model_kwargs
|
**model_kwargs
|
||||||
|
@ -12,6 +12,7 @@ def parse_args():
|
|||||||
default="lama",
|
default="lama",
|
||||||
choices=["lama", "ldm", "zits", "mat", "fcf", "sd1.5", "cv2", "manga", "sd2", "paint_by_example"],
|
choices=["lama", "ldm", "zits", "mat", "fcf", "sd1.5", "cv2", "manga", "sd2", "paint_by_example"],
|
||||||
)
|
)
|
||||||
|
parser.add_argument("--no-half", action="store_true", help="SD/PaintByExample model no half precision")
|
||||||
parser.add_argument(
|
parser.add_argument(
|
||||||
"--hf_access_token",
|
"--hf_access_token",
|
||||||
default="",
|
default="",
|
||||||
|
@ -303,6 +303,7 @@ def main(args):
|
|||||||
model = ModelManager(
|
model = ModelManager(
|
||||||
name=args.model,
|
name=args.model,
|
||||||
device=device,
|
device=device,
|
||||||
|
no_half=args.no_half,
|
||||||
hf_access_token=args.hf_access_token,
|
hf_access_token=args.hf_access_token,
|
||||||
sd_disable_nsfw=args.sd_disable_nsfw,
|
sd_disable_nsfw=args.sd_disable_nsfw,
|
||||||
sd_cpu_textencoder=args.sd_cpu_textencoder,
|
sd_cpu_textencoder=args.sd_cpu_textencoder,
|
||||||
|
Loading…
Reference in New Issue
Block a user