llama.cpp error
#1
by
mgonzs13
- opened
I have been using the GGUF version and since the commit https://github.com/ggerganov/llama.cpp/commit/f445c0e68cf8e1faca0b2aa8dfb9d48231cec301, I am getting GGML_ASSERT: llama.cpp:5533: n_embd_head == hparams.n_rot. Maybe regenerating the file can fix it.
Thanks for the heads up @mgonzs13 ! Rebuilding and updating shortly
@mgonzs13 Updated and tested with latest release (https://github.com/ggerganov/llama.cpp/commit/de473f5f8e19ba5e659cdf5af65fb9251dce16c5) 👍
salma-remyx
changed discussion status to
closed