ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Added k_norm & q_norm to merged Qwen3 layers (#2452)

Browse source
This commit is contained in:
cblomert 2025-05-02 12:07:37 +02:00 • committed by GitHub
commit 2ea34d47d5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 1 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

1
unsloth/save.py
View file

@ -70,6 +70,7 @@ LLAMA_WEIGHTS = (
LLAMA_LAYERNORMS = (
"input_layernorm", "post_attention_layernorm",
"pre_feedforward_layernorm", "post_feedforward_layernorm",
"self_attn.q_norm", "self_attn.k_norm",
)
# https://github.com/ggerganov/llama.cpp/blob/master/examples/quantize/quantize.cpp#L19