The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Cleanup. (#9838)
comfyanonymous29bf807b0e2d
diff --git a/comfy/audio_encoders/audio_encoders.py b/comfy/audio_encoders/audio_encoders.pyindex d1ec78f69..6fb5b08e9 100644--- a/comfy/audio_encoders/audio_encoders.py+++ b/comfy/audio_encoders/audio_encoders.py@@ -31,7 +31,7 @@ class AudioEncoderModel():def encode_audio(self, audio, sample_rate):comfy.model_management.load_model_gpu(self.patcher)audio = torchaudio.functional.resample(audio, sample_rate, self.model_sample_rate)- out, all_layers = self.model(audio.to(self.load_device), sr=self.model_sample_rate)+ out, all_layers = self.model(audio.to(self.load_device))outputs = {}outputs["encoded_audio"] = outoutputs["encoded_audio_all_layers"] = all_layersdiff --git a/comfy/audio_encoders/wav2vec2.py b/comfy/audio_encoders/wav2vec2.pyindex ef10dcd2a..4e34a40a7 100644--- a/comfy/audio_encoders/wav2vec2.py+++ b/comfy/audio_encoders/wav2vec2.py@@ -238,7 +238,7 @@ class Wav2Vec2Model(nn.Module):device=device, dtype=dtype, operations=operations)- def forward(self, x, sr=16000, mask_time_indices=None, return_dict=False):+ def forward(self, x, mask_time_indices=None, return_dict=False):x = torch.mean(x, dim=1)if self.do_normalize: