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

fix setting clamp_qkv value in OLMo conversion

Browse source
This commit is contained in:
nopperl 2024-04-25 17:34:20 +02:00
parent 853d06ffe2
commit 00f3fb6bc2
1 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
convert-hf-to-gguf.py
View file

@ -2764,8 +2764,9 @@ class OlmoModel(Model):
def set_gguf_parameters(self):
super().set_gguf_parameters()
self.gguf_writer.add_layer_norm_eps(1e-5)
if "clip_qkv" in self.hparams is not None:
self.gguf_writer.add_clamp_kqv(self.hparams["clip_qkv"])
clip_qkv = self.hparams.get("clip_qkv")
if clip_qkv is not None:
self.gguf_writer.add_clamp_kqv(clip_qkv)
# Same as super class, but permuting q_proj, k_proj
# Copied from: LlamaModel