298977887/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2025-04-14 23:53:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

Allow disabling pe in flux code for some other models.

Browse Source
This commit is contained in:
comfyanonymous 2025-03-18 05:09:25 -04:00
parent 50614f1b79
commit 3b19fc76e3
2 changed files with 10 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
comfy/ldm/flux/math.py
View File

@ -10,10 +10,11 @@ def attention(q: Tensor, k: Tensor, v: Tensor, pe: Tensor, mask=None) -> Tensor:
q_shape = q.shape q_shape = q.shape
k_shape = k.shape k_shape = k.shape
q = q.to(dtype=pe.dtype).reshape(*q.shape[:-1], -1, 1, 2) if pe is not None:
k = k.to(dtype=pe.dtype).reshape(*k.shape[:-1], -1, 1, 2) q = q.to(dtype=pe.dtype).reshape(*q.shape[:-1], -1, 1, 2)
q = (pe[..., 0] * q[..., 0] + pe[..., 1] * q[..., 1]).reshape(*q_shape).type_as(v) k = k.to(dtype=pe.dtype).reshape(*k.shape[:-1], -1, 1, 2)
k = (pe[..., 0] * k[..., 0] + pe[..., 1] * k[..., 1]).reshape(*k_shape).type_as(v) q = (pe[..., 0] * q[..., 0] + pe[..., 1] * q[..., 1]).reshape(*q_shape).type_as(v)
k = (pe[..., 0] * k[..., 0] + pe[..., 1] * k[..., 1]).reshape(*k_shape).type_as(v)
heads = q.shape[1] heads = q.shape[1]
x = optimized_attention(q, k, v, heads, skip_reshape=True, mask=mask) x = optimized_attention(q, k, v, heads, skip_reshape=True, mask=mask)

7
comfy/ldm/flux/model.py
View File

@ -115,8 +115,11 @@ class Flux(nn.Module):
vec = vec + self.vector_in(y[:,:self.params.vec_in_dim]) vec = vec + self.vector_in(y[:,:self.params.vec_in_dim])
txt = self.txt_in(txt) txt = self.txt_in(txt)
ids = torch.cat((txt_ids, img_ids), dim=1) if img_ids is not None:
pe = self.pe_embedder(ids) ids = torch.cat((txt_ids, img_ids), dim=1)
pe = self.pe_embedder(ids)
else:
pe = None
blocks_replace = patches_replace.get("dit", {}) blocks_replace = patches_replace.get("dit", {})
for i, block in enumerate(self.double_blocks): for i, block in enumerate(self.double_blocks):