vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

convert-hf : fix Refact conversion

Browse source
This commit is contained in:
Francis Couture-Harpin 2024-05-04 23:55:42 -04:00
parent f2099c50ab
commit 215a0d38c8
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
convert-hf-to-gguf.py
View file

@ -999,7 +999,7 @@ class RefactModel(Model):
ff_dim = multiple_of * ((hidden_dim + multiple_of - 1) // multiple_of) ff_dim = multiple_of * ((hidden_dim + multiple_of - 1) // multiple_of)
n_head = self.hparams["n_head"] n_head = self.hparams["n_head"]
n_head_kv = 1 n_head_kv = 1
head_dim = hidden_dim // n_head head_dim = self.hparams["n_embd"] // n_head
tensors: list[tuple[str, Tensor]] = [] tensors: list[tuple[str, Tensor]] = []