mradermacher

/

model_requests

like 84

Sorry for the late replay. The model failed and we likely forgot it was even requested by someone. I now forcefully requeued it to see why. Apparently there is an error loading the GGUF in llama.cpp due to token_embd.weight having the wrong shape.

llama_model_load: error loading model: check_tensor_dims: tensor 'token_embd.weight' has wrong shape; expected  5120, 131074, got  5120, 131072,     1,     1

While static quants started to upload I'm almost certain llama.cpp would not be able to load them so I decided to cancle the job and delete all the repositories using llmc nukeall AngelSlayer-Slush-12B