Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove no longer used code in the flux denoise function

Browse Source
This commit is contained in:
Brandon Rising 2024-08-26 13:07:31 -04:00 committed by Brandon
parent 3ea6c9666e
commit 849da67cc7
2 changed files with 3 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
invokeai/app/invocations/flux_text_to_image.py
View File

@ -112,7 +112,7 @@ class FluxTextToImageInvocation(BaseInvocation, WithMetadata, WithBoard):
with transformer_info as transformer: with transformer_info as transformer:
assert isinstance(transformer, Flux) assert isinstance(transformer, Flux)
def step_callback(img: torch.Tensor, state: PipelineIntermediateState) -> None: def step_callback() -> None:
if context.util.is_canceled(): if context.util.is_canceled():
raise CanceledException raise CanceledException

15
invokeai/backend/flux/sampling.py
View File

@ -109,7 +109,7 @@ def denoise(
vec: Tensor, vec: Tensor,
# sampling parameters # sampling parameters
timesteps: list[float], timesteps: list[float],
step_callback: Callable[[Tensor, PipelineIntermediateState], None], step_callback: Callable[[], None],
guidance: float = 4.0, guidance: float = 4.0,
): ):
dtype = model.txt_in.bias.dtype dtype = model.txt_in.bias.dtype
@ -123,7 +123,6 @@ def denoise(
# this is ignored for schnell # this is ignored for schnell
guidance_vec = torch.full((img.shape[0],), guidance, device=img.device, dtype=img.dtype) guidance_vec = torch.full((img.shape[0],), guidance, device=img.device, dtype=img.dtype)
step_count = 0
for t_curr, t_prev in tqdm(list(zip(timesteps[:-1], timesteps[1:], strict=True))): for t_curr, t_prev in tqdm(list(zip(timesteps[:-1], timesteps[1:], strict=True))):
t_vec = torch.full((img.shape[0],), t_curr, dtype=img.dtype, device=img.device) t_vec = torch.full((img.shape[0],), t_curr, dtype=img.dtype, device=img.device)
pred = model( pred = model(
@ -137,17 +136,7 @@ def denoise(
) )
img = img + (t_prev - t_curr) * pred img = img + (t_prev - t_curr) * pred
step_callback( step_callback()
img,
PipelineIntermediateState(
step=step_count,
order=0,
total_steps=len(timesteps),
timestep=math.floor(t_curr),
latents=img,
),
)
step_count += 1
return img return img