Skip to content

Fix Flux tuning issue #936

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

Merged
merged 7 commits into from
Oct 23, 2025
Merged

Fix Flux tuning issue #936

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
03fa925
Fix Flux tuning issue
mengniwang95 Oct 23, 2025
8f28568
Merge branch 'main' into mengni/flux_fix
mengniwang95 Oct 23, 2025
ea67bdb
Update base.py
mengniwang95 Oct 23, 2025
d648ce4
Update base.py
mengniwang95 Oct 23, 2025
e928702
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Oct 23, 2025
83aa797
Update base.py
mengniwang95 Oct 23, 2025
8dd7f24
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Oct 23, 2025
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
10 changes: 7 additions & 3 deletions auto_round/compressors/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2527,9 +2527,9 @@ def _get_current_num_elm(
def _quantize_block(
self,
block: torch.nn.Module,
input_ids: list[torch.Tensor],
input_ids: Union[list[torch.Tensor], dict],
input_others: dict,
q_input: Union[None, torch.Tensor] = None,
q_input: Union[torch.Tensor, dict, None] = None,
device: Union[str, torch.device] = "cpu",
):
"""Quantize the weights of a given block of the model.
Expand Down Expand Up @@ -2646,7 +2646,11 @@ def _quantize_block(
else:
lr_schedule = copy.deepcopy(self.lr_scheduler)

nsamples = len(input_ids)
if isinstance(input_ids, dict): # input_ids of Flux is dict
nsamples = len(input_ids["hidden_states"])
else:
nsamples = len(input_ids)

pick_samples = self.batch_size * self.gradient_accumulate_steps
pick_samples = min(nsamples, pick_samples)
if self.sampler != "rand":
Expand Down
7 changes: 5 additions & 2 deletions auto_round/compressors/diffusion/compressor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -210,7 +210,7 @@ def _get_current_q_output(
def _get_block_outputs(
self,
block: torch.nn.Module,
input_ids: torch.Tensor,
input_ids: Union[torch.Tensor, dict],
input_others: torch.Tensor,
bs: int,
device: Union[str, torch.device],
Expand All @@ -233,8 +233,11 @@ def _get_block_outputs(
"""

output = defaultdict(list)
nsamples = len(input_ids)
output_config = output_configs.get(block.__class__.__name__, [])
if isinstance(input_ids, dict):
nsamples = len(input_ids["hidden_states"])
else:
nsamples = len(input_ids)

for i in range(0, nsamples, bs):
end_index = min(nsamples, i + bs)
Expand Down