Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature/fixed rand generator #25

Merged
merged 5 commits into from
Apr 24, 2023
Merged

Feature/fixed rand generator #25

Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
2959c29
Added random seed argument.
alihassanijr Apr 23, 2023
2c81205
Add generator to the module, otherwise it's the same set of random nu…
alihassanijr Apr 24, 2023
0d1b978
Pull generator out of tome_info?
alihassanijr Apr 24, 2023
bed066d
Back to one generator
alihassanijr Apr 24, 2023
5453a71
Init generator in forward.
alihassanijr Apr 24, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Init generator in forward.
  • Loading branch information
@alihassanijr
alihassanijr committed Apr 24, 2023
commit 5453a718b0bb26aed104e7d27396a7ff4af42ee4
24 changes: 17 additions & 7 deletions tomesd/patch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,20 @@



def init_generator(device: torch.device):
"""
Forks the current default random generator given device.
"""
if device.type == "cpu":
return torch.Generator(device=device).set_state(torch.get_rng_state())
elif device.type == "cuda":
return torch.Generator(device=device).set_state(torch.cuda.get_rng_state())
elif device.type == "mps":
return torch.Generator(device=device).set_state(torch.mps.get_rng_state())
raise NotImplementedError(f"Invalid/unsupported device. Expected `cpu`, `cuda`, or `mps`, got {device.type}.")



def compute_merge(x: torch.Tensor, tome_info: Dict[str, Any]) -> Tuple[Callable, ...]:
original_h, original_w = tome_info["size"]
original_tokens = original_h * original_w
Expand All @@ -20,6 +34,8 @@ def compute_merge(x: torch.Tensor, tome_info: Dict[str, Any]) -> Tuple[Callable,
r = int(x.shape[1] * args["ratio"])
# If the batch size is odd, then it's not possible for prompted and unprompted images to be in the same
# batch, which causes artifacts with use_rand, so force it to be off.
if args["generator"] is None or args["generator"].device != x.device:
args["generator"] = init_generator(x.device)
use_rand = False if x.shape[0] % 2 == 1 else args["use_rand"]
generator = None if x.shape[0] % 2 == 1 else args["generator"]
m, u = merge.bipartite_soft_matching_random2d(x, w, h, args["sx"], args["sy"], r,
Expand Down Expand Up @@ -178,7 +194,6 @@ def apply_patch(
max_downsample: int = 1,
sx: int = 2, sy: int = 2,
use_rand: bool = True,
rand_seed: int = None,
merge_attn: bool = True,
merge_crossattn: bool = False,
merge_mlp: bool = False):
Expand Down Expand Up @@ -219,11 +234,6 @@ def apply_patch(
# Supports "pipe.unet" and "unet"
diffusion_model = model.unet if hasattr(model, "unet") else model

generator = None
if rand_seed is not None:
generator = torch.Generator(device=diffusion_model.device)
generator = generator.manual_seed(rand_seed)

diffusion_model._tome_info = {
"size": None,
"hooks": [],
Expand All @@ -232,7 +242,7 @@ def apply_patch(
"max_downsample": max_downsample,
"sx": sx, "sy": sy,
"use_rand": use_rand,
"generator": generator,
"generator": None,
"merge_attn": merge_attn,
"merge_crossattn": merge_crossattn,
"merge_mlp": merge_mlp
Expand Down