llama.cpp
to allow loading of karpathy/tinyllamas pt weights, assume model weights in 'model' value if exists
#4089
Merged

Loading