FireRed-ImageEdit-Japanese

Running on Zero

someone-in-the-world Claude Sonnet 4.6 commited on May 22

Commit

f4a7c5c

1 Parent(s): 519548d

Raise high-detail ZeroGPU duration to 90s (was 60s)

Latest 100 logs show high-detail p95=67.9s and p99=72.6s pipeline time,
so 60s was insufficient. 90s covers p99 with headroom for overhead.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Files changed (1) hide show

app.py +1 -1

app.py CHANGED Viewed

@@ -352,7 +352,7 @@ def infer(images_b64_json, prompt, seed, randomize_seed, guidance_scale, steps,
         raise
-@spaces.GPU(duration=lambda *a, **kw: 30 if (len(a) > 7 and a[7] == "fast") else 60)
 def _infer_gpu(pil_images, prompt, seed, guidance_scale, steps, width, height, mode):
     _cuda_ok = torch.cuda.is_available()
     timer = _InferTimer(_cuda_ok)

         raise
+@spaces.GPU(duration=lambda *a, **kw: 30 if (len(a) > 7 and a[7] == "fast") else 90)
 def _infer_gpu(pil_images, prompt, seed, guidance_scale, steps, width, height, mode):
     _cuda_ok = torch.cuda.is_available()
     timer = _InferTimer(_cuda_ok)