Skip to content

fix: add ACESTEP_SKIP_VRAM_PREFLIGHT env var + gc before preflight check #1091

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

Open
FlexOr2 wants to merge 3 commits into
base: main
Choose a base branch
from
Open

fix: add ACESTEP_SKIP_VRAM_PREFLIGHT env var + gc before preflight check #1091

Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
84f3a08
fix: add ACESTEP_SKIP_VRAM_PREFLIGHT env var + gc before check
FlexOr2 Apr 12, 2026
5166c17
fix: raise VRAM preflight skip notice from debug to warning
FlexOr2 Apr 19, 2026
d67c4a6
fix: gate VRAM preflight cleanup/skip-warning on CUDA availability
FlexOr2 Apr 21, 2026
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
27 changes: 20 additions & 7 deletions acestep/core/generation/handler/generate_music.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@
"""

import gc
import os
import traceback
from typing import Any, Dict, List, Optional, Union

Expand Down Expand Up @@ -328,13 +329,25 @@ def generate_music(
repainting_end=repainting_end,
chunk_mask_mode=chunk_mask_mode,
)
vram_error = self._vram_preflight_check(
actual_batch_size=actual_batch_size,
audio_duration=audio_duration,
guidance_scale=guidance_scale,
)
if vram_error is not None:
return vram_error
gc.collect()
torch.cuda.empty_cache()
skip_preflight = os.environ.get(
"ACESTEP_SKIP_VRAM_PREFLIGHT", "",
).lower() in ("1", "true", "yes")
if skip_preflight:
logger.warning(
"[generate_music] VRAM pre-flight check skipped via "
"ACESTEP_SKIP_VRAM_PREFLIGHT=1. If generation OOMs, "
"unset this variable to re-enable the safety check."
)
else:
vram_error = self._vram_preflight_check(
actual_batch_size=actual_batch_size,
audio_duration=audio_duration,
guidance_scale=guidance_scale,
)
if vram_error is not None:
return vram_error
Comment thread
coderabbitai[bot] marked this conversation as resolved.
Outdated

injection_ratio, resolved_cf_frames, resolved_wav_cf = (
_resolve_repaint_config(repaint_mode, repaint_strength)
Expand Down