Update app.py

app.py

@@ -537,7 +537,21 @@ def _generate(
             video = pk.interpolate(FILM, video, multiplier)
     fps = FPS * multiplier
 
-    return video, audio, sampling_rate, fps, multiplier
+    # Muxed to an mp4 here, before returning, rather than in the caller: a raw CUDA tensor can't cross a
+    # `@spaces.GPU` return at all under ZeroGPU's CUDA-emulation mode (`RuntimeError: Low-level CUDA init
+    # reached` trying to reconstruct it in the dispatching process), and a CPU float tensor of several hundred
+    # interpolated frames is needlessly large to pickle anyway when the finished file is a few MB of h264.
+    from diffusers.utils import encode_video
+
+    frames = (video.permute(0, 2, 3, 1).float() * 255.0).round_().clamp_(0, 255).to(torch.uint8).cpu()
+    del video
+
+    directory = os.path.join(tempfile.gettempdir(), "h3-outputs")
+    os.makedirs(directory, exist_ok=True)
+    path = os.path.join(directory, f"h3-ref2va-{int(time.time() * 1000)}.mp4")
+    encode_video(frames, fps=fps, output_path=path, audio=audio, audio_sample_rate=sampling_rate)
+
+    return path, fps, multiplier
 
 
 def generate(
@@ -609,17 +623,12 @@ def generate(
 
     progress(0.1, desc=f"Generating {num_frames / FPS:.1f} s at {width}x{height} ...")
     started = time.time()
-    video, audio, sampling_rate, fps, multiplier = _generate(
+    path, fps, multiplier = _generate(
         prompt_embeds, text_token_tags, references, height, width, num_frames, steps, seed,
         sampler_key, schedule_key, float(video_shift), float(audio_shift), float(sharpen), multiplier,
     )
     generate_seconds = time.time() - started
 
-    directory = os.path.join(tempfile.gettempdir(), "h3-outputs")
-    os.makedirs(directory, exist_ok=True)
-    path = os.path.join(directory, f"h3-ref2va-{int(time.time() * 1000)}.mp4")
-    encode_video(video, fps=fps, output_path=path, audio=audio, audio_sample_rate=sampling_rate)
-
     print(
         f"[ref2va] {[kind for kind, _ in references]} · `{width}x{height}`, {num_frames} frames "
         f"({num_frames / FPS:.3f} s), {int(steps)} steps of `{schedule_key}` · sampler `{sampler_key}` · "