llama.cpp
llama : gemma3 : use output tensor if it exists in model weight
#12506
Merged
Merged
llama : gemma3 : use output tensor if it exists in model weight #12506
ngxson
merged
2 commits into
ggml-org:master
from
ngxson:xsn/gemma3_lm_head
llama : gemma3 : use output tensor if it exists in model weight
9a49ae9f
also add to the llm_tensor_names
ab0ff0f7
ggerganov
approved these changes
on 2025-03-22
Login to write a write a comment.
Login via GitHub