Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

save VRAM by not recombining tensors that have been sliced to save VRAM

Browse Source
This commit is contained in:
damian0815 2022-11-02 00:31:58 +01:00 committed by blessedcoolant
parent dff5681cf0
commit 4513320bf1
1 changed files with 19 additions and 58 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

73
ldm/models/diffusion/cross_attention_control.py
View File

@ -151,72 +151,33 @@ class CrossAttentionControl:
#else: #else:
# print(f"in wrangler, whole, use_last_attn_slice is {self.use_last_attn_slice}, save_last_attn_slice is {self.save_last_attn_slice}") # print(f"in wrangler, whole, use_last_attn_slice is {self.use_last_attn_slice}, save_last_attn_slice is {self.save_last_attn_slice}")
if self.use_last_attn_slice: if self.use_last_attn_slice:
this_attn_slice = attn_slice
if self.last_attn_slice_mask is not None:
# indices and mask operate on dim=2, no need to slice
base_attn_slice_full = torch.index_select(self.last_attn_slice, -1, self.last_attn_slice_indices)
base_attn_slice_mask = self.last_attn_slice_mask
if dim is None: if dim is None:
base_attn_slice = base_attn_slice_full last_attn_slice = self.last_attn_slice
#print("using whole base slice of shape", base_attn_slice.shape, "from complete shape", base_attn_slice_full.shape) # print("took whole slice of shape", attn_slice.shape, "from complete shape", self.last_attn_slice.shape)
elif dim == 0:
base_attn_slice = base_attn_slice_full[start:end]
#print("using base dim 0 slice of shape", base_attn_slice.shape, "from complete shape", base_attn_slice_full.shape)
elif dim == 1:
base_attn_slice = base_attn_slice_full[:, start:end]
#print("using base dim 1 slice of shape", base_attn_slice.shape, "from complete shape", base_attn_slice_full.shape)
attn_slice = this_attn_slice * (1 - base_attn_slice_mask) + \
base_attn_slice * base_attn_slice_mask
else: else:
if dim is None: last_attn_slice = self.last_attn_slice[offset]
attn_slice = self.last_attn_slice
#print("took whole slice of shape", attn_slice.shape, "from complete shape", self.last_attn_slice.shape) if self.last_attn_slice_mask is None:
elif dim == 0: # just use everything
attn_slice = self.last_attn_slice[start:end] attn_slice = last_attn_slice
#print("took dim 0 slice of shape", attn_slice.shape, "from complete shape", self.last_attn_slice.shape) else:
elif dim == 1: last_attn_slice_mask = self.last_attn_slice_mask
attn_slice = self.last_attn_slice[:, start:end] remapped_last_attn_slice = torch.index_select(last_attn_slice, -1, self.last_attn_slice_indices)
#print("took dim 1 slice of shape", attn_slice.shape, "from complete shape", self.last_attn_slice.shape)
this_attn_slice = attn_slice
this_attn_slice_mask = 1 - last_attn_slice_mask
attn_slice = this_attn_slice * this_attn_slice_mask + \
remapped_last_attn_slice * last_attn_slice_mask
if self.save_last_attn_slice: if self.save_last_attn_slice:
if dim is None: if dim is None:
self.last_attn_slice = attn_slice self.last_attn_slice = attn_slice
elif dim == 0:
# dynamically grow last_attn_slice if needed
if self.last_attn_slice is None:
self.last_attn_slice = attn_slice
#print("no last_attn_slice: shape now", self.last_attn_slice.shape)
elif self.last_attn_slice.shape[0] == start:
self.last_attn_slice = torch.cat([self.last_attn_slice, attn_slice], dim=0)
assert(self.last_attn_slice.shape[0] == end)
#print("last_attn_slice too small, appended dim 0 shape", attn_slice.shape, ", shape now", self.last_attn_slice.shape)
else: else:
# no need to grow
self.last_attn_slice[start:end] = attn_slice
#print("last_attn_slice shape is fine, setting dim 0 shape", attn_slice.shape, ", shape now", self.last_attn_slice.shape)
elif dim == 1:
# dynamically grow last_attn_slice if needed
if self.last_attn_slice is None: if self.last_attn_slice is None:
self.last_attn_slice = attn_slice self.last_attn_slice = { offset: attn_slice }
elif self.last_attn_slice.shape[1] == start:
self.last_attn_slice = torch.cat([self.last_attn_slice, attn_slice], dim=1)
assert(self.last_attn_slice.shape[1] == end)
else: else:
# no need to grow self.last_attn_slice[offset] = attn_slice
self.last_attn_slice[:, start:end] = attn_slice
if self.use_last_attn_weights and self.last_attn_slice_weights is not None:
if dim is None:
weights = self.last_attn_slice_weights
elif dim == 0:
weights = self.last_attn_slice_weights[start:end]
elif dim == 1:
weights = self.last_attn_slice_weights[:, start:end]
attn_slice = attn_slice * weights
return attn_slice return attn_slice