Update app.py

app.py

@@ -166,24 +166,6 @@ class LTX23DistilledA2VPipeline:
 
         video_duration = num_frames / frame_rate
 
-        encoded_audio_latent = vae_encode_audio(self.model_ledger.audio_encoder())
-        audio_shape = AudioLatentShape.from_duration(batch=1, duration=video_duration, channels=8, mel_bins=16)
-        expected_frames = audio_shape.frames
-        actual_frames = encoded_audio_latent.shape[2]
-
-        if actual_frames > expected_frames:
-            encoded_audio_latent = encoded_audio_latent[:, :, :expected_frames, :]
-        elif actual_frames < expected_frames:
-            pad = torch.zeros(
-                encoded_audio_latent.shape[0],
-                encoded_audio_latent.shape[1],
-                expected_frames - actual_frames,
-                encoded_audio_latent.shape[3],
-                device=encoded_audio_latent.device,
-                dtype=encoded_audio_latent.dtype,
-            )
-            encoded_audio_latent = torch.cat([encoded_audio_latent, pad], dim=2)
-
         video_encoder = self.model_ledger.video_encoder()
         transformer = self.model_ledger.transformer()
         stage_1_sigmas = torch.tensor(DISTILLED_SIGMA_VALUES, device=self.device)