llama.cpp
Fix kq_scale for the attention layers of PLaMo2
#14892

Merged

Commits

Fix dimensions for expand

mitmul committed 335 days ago
Change dimensions to copy states to cache

mitmul committed 327 days ago
Fix the default value for plamo2 conversion

mitmul committed 327 days ago
Fix scale given to build_attn

mitmul committed 327 days ago
Update src/llama-model.cpp

mitmul committed 326 days ago
Update src/llama-model.cpp

mitmul committed 326 days ago
Update src/llama-model.cpp

mitmul committed 326 days ago

Loading