techfreakworm commited on
Commit
0a98b65
·
unverified ·
1 Parent(s): 10af3aa

fix(spaces): _duration_for must accept progress kwarg to match _execute_workflow

Browse files

ZeroGPU calls the duration estimator with the same args it'll pass to the
@spaces.GPU function. After adding progress as the 7th arg to
_execute_workflow in 27ffbbc, _duration_for raised TypeError on every
generation. Add progress to its signature; we ignore it for estimation.

Files changed (1) hide show
  1. backend.py +5 -0
backend.py CHANGED
@@ -99,9 +99,14 @@ def _duration_for(
99
  mode: str,
100
  preset: str,
101
  multiplier: float = 1.0,
 
102
  ) -> int:
103
  """ZeroGPU duration estimator. Same signature as _execute_workflow.
104
 
 
 
 
 
105
  Estimate = (base × preset multiplier + cold-cache buffer + per-frame VAE
106
  decode time) × retry multiplier, clamped to [60s, 900s]. The 900s ceiling
107
  keeps a single failed call from torching the daily quota.
 
99
  mode: str,
100
  preset: str,
101
  multiplier: float = 1.0,
102
+ progress: Any = None,
103
  ) -> int:
104
  """ZeroGPU duration estimator. Same signature as _execute_workflow.
105
 
106
+ `progress` is a gr.Progress instance forwarded by the caller; we ignore it
107
+ here (estimator doesn't emit progress) but must accept it positionally so
108
+ ZeroGPU can call us with the same arg list it'll use for _execute_workflow.
109
+
110
  Estimate = (base × preset multiplier + cold-cache buffer + per-frame VAE
111
  decode time) × retry multiplier, clamped to [60s, 900s]. The 900s ceiling
112
  keeps a single failed call from torching the daily quota.