wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-01-10 06:10:50 +08:00
Code Issues Actions 10 Packages Projects Releases Wiki Activity

audio_encoders now a package correctly, make imports relative

Browse Source
This commit is contained in:
doctorpangloss 2025-08-26 14:12:40 -07:00
parent 752caf013c
commit 306fcbaa0e
2 changed files with 12 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

0
comfy/audio_encoders/__init__.py Normal file
View File

23
comfy/audio_encoders/audio_encoders.py
View File

@ -1,17 +1,16 @@
from .wav2vec2 import Wav2Vec2Model
import comfy.model_management
import comfy.ops
import comfy.utils
import logging
import torchaudio
from ..model_management import text_encoder_offload_device, text_encoder_device, load_model_gpu, text_encoder_dtype
from ..ops import manual_cast
from ..utils import state_dict_prefix_replace
import logging
class AudioEncoderModel():
def __init__(self, config):
self.load_device = comfy.model_management.text_encoder_device()
offload_device = comfy.model_management.text_encoder_offload_device()
self.dtype = comfy.model_management.text_encoder_dtype(self.load_device)
self.model = Wav2Vec2Model(dtype=self.dtype, device=offload_device, operations=comfy.ops.manual_cast)
self.load_device = text_encoder_device()
offload_device = text_encoder_offload_device()
self.dtype = text_encoder_dtype(self.load_device)
self.model = Wav2Vec2Model(dtype=self.dtype, device=offload_device, operations=manual_cast)
self.model.eval()
self.patcher = comfy.model_patcher.ModelPatcher(self.model, load_device=self.load_device, offload_device=offload_device)
self.model_sample_rate = 16000
@ -23,7 +22,9 @@ class AudioEncoderModel():
return self.model.state_dict()
def encode_audio(self, audio, sample_rate):
comfy.model_management.load_model_gpu(self.patcher)
# this one we will allow to just bubble up the exception
import torchaudio # pylint: disable=import-error
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))
outputs = {}
@ -34,7 +35,7 @@ class AudioEncoderModel():
def load_audio_encoder_from_sd(sd, prefix=""):
audio_encoder = AudioEncoderModel(None)
sd = comfy.utils.state_dict_prefix_replace(sd, {"wav2vec2.": ""})
sd = state_dict_prefix_replace(sd, {"wav2vec2.": ""})
m, u = audio_encoder.load_sd(sd)
if len(m) > 0:
logging.warning("missing audio encoder: {}".format(m))