jaidaken/ComfyUI
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Basic support for the ace step 1.5 model. (#12237)

Browse Source
This commit is contained in:
comfyanonymous
2026-02-02 21:06:18 -08:00
committed by GitHub
parent ba5bf3f1a8
commit 3c1a1a2df8
12 changed files with 1566 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
comfy/sd.py
View File

@@ -59,6 +59,7 @@ import comfy.text_encoders.kandinsky5
import comfy.text_encoders.jina_clip_2
import comfy.text_encoders.newbie
import comfy.text_encoders.anima
import comfy.text_encoders.ace15
import comfy.model_patcher
import comfy.lora
@@ -452,6 +453,8 @@ class VAE:
self.extra_1d_channel = None
self.crop_input = True
self.audio_sample_rate = 44100
if config is None:
if "decoder.mid.block_1.mix_factor" in sd:
encoder_config = {'double_z': True, 'z_channels': 4, 'resolution': 256, 'in_channels': 3, 'out_ch': 3, 'ch': 128, 'ch_mult': [1, 2, 4, 4], 'num_res_blocks': 2, 'attn_resolutions': [], 'dropout': 0.0}
@@ -549,14 +552,25 @@ class VAE:
encoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Encoder", 'params': ddconfig},
decoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Decoder", 'params': ddconfig})
elif "decoder.layers.1.layers.0.beta" in sd:
self.first_stage_model = AudioOobleckVAE()
config = {}
param_key = None
if "decoder.layers.2.layers.1.weight_v" in sd:
param_key = "decoder.layers.2.layers.1.weight_v"
if "decoder.layers.2.layers.1.parametrizations.weight.original1" in sd:
param_key = "decoder.layers.2.layers.1.parametrizations.weight.original1"
if param_key is not None:
if sd[param_key].shape[-1] == 12:
config["strides"] = [2, 4, 4, 6, 10]
self.audio_sample_rate = 48000
self.first_stage_model = AudioOobleckVAE(**config)
self.memory_used_encode = lambda shape, dtype: (1000 * shape[2]) * model_management.dtype_size(dtype)
self.memory_used_decode = lambda shape, dtype: (1000 * shape[2] * 2048) * model_management.dtype_size(dtype)
self.latent_channels = 64
self.output_channels = 2
self.pad_channel_value = "replicate"
self.upscale_ratio = 2048
self.downscale_ratio = 2048
self.downscale_ratio = 2048
self.latent_dim = 1
self.process_output = lambda audio: audio
self.process_input = lambda audio: audio
@@ -1427,6 +1441,9 @@ def load_text_encoder_state_dicts(state_dicts=[], embedding_directory=None, clip
clip_data_jina = clip_data[0]
tokenizer_data["gemma_spiece_model"] = clip_data_gemma.get("spiece_model", None)
tokenizer_data["jina_spiece_model"] = clip_data_jina.get("spiece_model", None)
elif clip_type == CLIPType.ACE:
clip_target.clip = comfy.text_encoders.ace15.te(**llama_detect(clip_data))
clip_target.tokenizer = comfy.text_encoders.ace15.ACE15Tokenizer
else:
clip_target.clip = sdxl_clip.SDXLClipModel
clip_target.tokenizer = sdxl_clip.SDXLTokenizer