298977887/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2025-04-19 19:03:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'comfyanonymous:master' into master

Browse Source
This commit is contained in:
Red 2025-03-20 04:25:30 -07:00 committed by GitHub
commit e562ca4728
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 15 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
comfy/ldm/hunyuan3d/vae.py
View File

@ -581,7 +581,7 @@ class ShapeVAE(nn.Module):
enable_pbar = kwargs.get("enable_pbar", True)
grid_logits = self.volume_decoder(latents, self.geo_decoder, bounds=bounds, num_chunks=num_chunks, octree_resolution=octree_resolution, enable_pbar=enable_pbar)
return grid_logits
return grid_logits.movedim(-2, -1)
def encode(self, x):
return None

5
comfy/sd.py
View File

@ -419,10 +419,11 @@ class VAE:
inner_size = sd["geo_decoder.output_proj.weight"].shape[1]
downsample_ratio = sd["post_kl.weight"].shape[0] // inner_size
mlp_expand = sd["geo_decoder.cross_attn_decoder.mlp.c_fc.weight"].shape[0] // inner_size
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.memory_used_encode = lambda shape, dtype: (1000 * shape[2]) * model_management.dtype_size(dtype) # TODO
self.memory_used_decode = lambda shape, dtype: (1024 * 1024 * 1024 * 2.0) * model_management.dtype_size(dtype) # TODO
ddconfig = {"embed_dim": 64, "num_freqs": 8, "include_pi": False, "heads": 16, "width": 1024, "num_decoder_layers": 16, "qkv_bias": False, "qk_norm": True, "geo_decoder_mlp_expand_ratio": mlp_expand, "geo_decoder_downsample_ratio": downsample_ratio, "geo_decoder_ln_post": ln_post}
self.first_stage_model = comfy.ldm.hunyuan3d.vae.ShapeVAE(**ddconfig)
self.working_dtypes = [torch.float16, torch.bfloat16, torch.float32]
else:
logging.warning("WARNING: No VAE weights detected, VAE not initalized.")
self.first_stage_model = None

2
comfy/supported_models.py
View File

@ -971,6 +971,8 @@ class Hunyuan3Dv2(supported_models_base.BASE):
"shift": 1.0,
}
memory_usage_factor = 3.5
clip_vision_prefix = "conditioner.main_image_encoder.model."
vae_key_prefix = ["vae."]

13
comfy_extras/nodes_hunyuan3d.py
View File

@ -185,13 +185,17 @@ def voxel_to_mesh(voxels, threshold=0.5, device=None):
device=device
).reshape(-1, 4)
all_indices.append(torch.stack([face_indices[:, 0], face_indices[:, 2], face_indices[:, 1]], dim=1))
all_indices.append(torch.stack([face_indices[:, 0], face_indices[:, 3], face_indices[:, 2]], dim=1))
all_indices.append(torch.stack([face_indices[:, 0], face_indices[:, 1], face_indices[:, 2]], dim=1))
all_indices.append(torch.stack([face_indices[:, 0], face_indices[:, 2], face_indices[:, 3]], dim=1))
vertex_count += 4 * num_faces
vertices = torch.cat(all_vertices, dim=0)
faces = torch.cat(all_indices, dim=0)
if len(all_vertices) > 0:
vertices = torch.cat(all_vertices, dim=0)
faces = torch.cat(all_indices, dim=0)
else:
vertices = torch.zeros((1, 3))
faces = torch.zeros((1, 3))
v_min = 0
v_max = max(voxels.shape)
@ -202,6 +206,7 @@ def voxel_to_mesh(voxels, threshold=0.5, device=None):
if scale > 0:
vertices = vertices / scale
vertices = torch.fliplr(vertices)
return vertices, faces