vrr/kvcache-ai-ktransformers
2
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2025-09-10 23:34:35 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #657 from kvcache-ai/feat-absorb-for-long-prefill

Browse source 
Feat absorb for long prefill
This commit is contained in:
Atream 2025-02-25 16:53:21 +08:00 • committed by GitHub
commit b443c7dfa2
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
11 changed files with 193 additions and 43 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ktransformers/util/custom_gguf.py
View file

@ -330,6 +330,8 @@ class GGUFLoader:
values = GGML_DEQUANTIZE[ggml_name](data)
values = torch.from_numpy(values.copy())
if ggml_name == "BF16":
values = values.view(torch.bfloat16)
values = values.view(shape[-2::-1])
return values