Giving me error with llama-cpp-python
#11
by
divyanshu-k
- opened
Hi, I downloaded a model GGUF to run it locally using the llama cpp python library. However, it shows that tensor 'blk.0.ffn_down_exps.weight' has invalid ggml type 39 (NONE) which means that the ggml version is not compatible with the library. Since I'm new to this, I would like if you could tell me how to run it locally.
You need to recompile llama.cpp from scratch and update it!