mirrors/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2025-09-10 19:46:38 +00:00
Code Issues Packages Projects Releases Wiki Activity

WIP support for Wan t2v model.

Browse Source
This commit is contained in:
comfyanonymous
2025-02-25 17:20:35 -05:00
parent f40076096e
commit 63023011b9
10 changed files with 1307 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
comfy/supported_models.py
View File

@@ -16,6 +16,7 @@ import comfy.text_encoders.lt
import comfy.text_encoders.hunyuan_video
import comfy.text_encoders.cosmos
import comfy.text_encoders.lumina2
import comfy.text_encoders.wan
from . import supported_models_base
from . import latent_formats
@@ -895,6 +896,35 @@ class Lumina2(supported_models_base.BASE):
hunyuan_detect = comfy.text_encoders.hunyuan_video.llama_detect(state_dict, "{}gemma2_2b.transformer.".format(pref))
return supported_models_base.ClipTarget(comfy.text_encoders.lumina2.LuminaTokenizer, comfy.text_encoders.lumina2.te(**hunyuan_detect))
models = [Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, FluxSchnell, GenmoMochi, LTXV, HunyuanVideo, CosmosT2V, CosmosI2V, Lumina2]
class WAN21_T2V(supported_models_base.BASE):
unet_config = {
"image_model": "wan2.1",
"model_type": "t2v",
}
sampling_settings = {
"shift": 8.0,
}
unet_extra_config = {}
latent_format = latent_formats.Wan21
memory_usage_factor = 1.0
supported_inference_dtypes = [torch.bfloat16, torch.float32]
vae_key_prefix = ["vae."]
text_encoder_key_prefix = ["text_encoders."]
def get_model(self, state_dict, prefix="", device=None):
out = model_base.WAN21_T2V(self, device=device)
return out
def clip_target(self, state_dict={}):
pref = self.text_encoder_key_prefix[0]
t5_detect = comfy.text_encoders.sd3_clip.t5_xxl_detect(state_dict, "{}umt5xxl.transformer.".format(pref))
return supported_models_base.ClipTarget(comfy.text_encoders.wan.WanT5Tokenizer, comfy.text_encoders.wan.te(**t5_detect))
models = [Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, FluxSchnell, GenmoMochi, LTXV, HunyuanVideo, CosmosT2V, CosmosI2V, Lumina2, WAN21_T2V]
models += [SVD_img2vid]