wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-04-23 08:52:32 +08:00
Code Issues Actions 23 Packages Projects Releases Wiki Activity

Support standalone LTXV audio VAEs (#13499)

Browse Source
This commit is contained in:
Jukka Seppänen 2026-04-21 20:46:37 +03:00 committed by GitHub
parent b38dd0ff23
commit eb22225387
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
comfy/sd.py
View File

@ -807,6 +807,7 @@ class VAE:
self.memory_used_encode = lambda shape, dtype: (700 * (max(1, (shape[-3] ** 0.66 * 0.11)) * shape[-2] * shape[-1]) * model_management.dtype_size(dtype)) self.memory_used_encode = lambda shape, dtype: (700 * (max(1, (shape[-3] ** 0.66 * 0.11)) * shape[-2] * shape[-1]) * model_management.dtype_size(dtype))
self.memory_used_decode = lambda shape, dtype: (50 * (max(1, (shape[-3] ** 0.65 * 0.26)) * shape[-2] * shape[-1] * 32 * 32) * model_management.dtype_size(dtype)) self.memory_used_decode = lambda shape, dtype: (50 * (max(1, (shape[-3] ** 0.65 * 0.26)) * shape[-2] * shape[-1] * 32 * 32) * model_management.dtype_size(dtype))
elif "vocoder.resblocks.0.convs1.0.weight" in sd or "vocoder.vocoder.resblocks.0.convs1.0.weight" in sd: # LTX Audio elif "vocoder.resblocks.0.convs1.0.weight" in sd or "vocoder.vocoder.resblocks.0.convs1.0.weight" in sd: # LTX Audio
sd = comfy.utils.state_dict_prefix_replace(sd, {"audio_vae.": "autoencoder."})
self.first_stage_model = comfy.ldm.lightricks.vae.audio_vae.AudioVAE(metadata=metadata) self.first_stage_model = comfy.ldm.lightricks.vae.audio_vae.AudioVAE(metadata=metadata)
self.memory_used_encode = lambda shape, dtype: (shape[2] * 330) * model_management.dtype_size(dtype) self.memory_used_encode = lambda shape, dtype: (shape[2] * 330) * model_management.dtype_size(dtype)
self.memory_used_decode = lambda shape, dtype: (shape[2] * shape[3] * 87000) * model_management.dtype_size(dtype) self.memory_used_decode = lambda shape, dtype: (shape[2] * shape[3] * 87000) * model_management.dtype_size(dtype)