lllyasviel/stable-diffusion-webui-forge
1
0
Fork 0
mirror of https://github.com/lllyasviel/stable-diffusion-webui-forge.git synced 2026-05-01 03:31:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

ling

Browse Source
This commit is contained in:
layerdiffusion
2024-08-15 00:33:21 -07:00
parent fb62214a32
commit 0524133714
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
backend/operations_gguf.py
View File

@@ -58,7 +58,7 @@ def dequantize_blocks_Q8_0(blocks, block_size, type_size):
d = blocks[:, :2].view(torch.float16) d = blocks[:, :2].view(torch.float16)
x = blocks[:, 2:].view(torch.int8).to(torch.float16) x = blocks[:, 2:].view(torch.int8).to(torch.float16)
return (x * d) return x * d
def dequantize_blocks_Q5_0(blocks, block_size, type_size): def dequantize_blocks_Q5_0(blocks, block_size, type_size):
@@ -80,7 +80,7 @@ def dequantize_blocks_Q5_0(blocks, block_size, type_size):
ql = (ql & 0x0F).reshape(n_blocks, -1) ql = (ql & 0x0F).reshape(n_blocks, -1)
qs = (ql | (qh << 4)).to(torch.int8) - 16 qs = (ql | (qh << 4)).to(torch.int8) - 16
return (d * qs) return d * qs
def dequantize_blocks_Q4_0(blocks, block_size, type_size): def dequantize_blocks_Q4_0(blocks, block_size, type_size):
@@ -93,7 +93,7 @@ def dequantize_blocks_Q4_0(blocks, block_size, type_size):
qs = qs.reshape((n_blocks, -1, 1, block_size // 2)) >> torch.tensor([0, 4], device=d.device, dtype=torch.uint8).reshape((1, 1, 2, 1)) qs = qs.reshape((n_blocks, -1, 1, block_size // 2)) >> torch.tensor([0, 4], device=d.device, dtype=torch.uint8).reshape((1, 1, 2, 1))
qs = (qs & 0x0F).reshape((n_blocks, -1)).to(torch.int8) - 8 qs = (qs & 0x0F).reshape((n_blocks, -1)).to(torch.int8) - 8
return (d * qs) return d * qs
dequantize_functions = { dequantize_functions = {