Failed to load model
#1
by Topsy1 - opened
Hello, im using the IQ4XS version and i keep getting this error
llama_model_load: error loading model: check_tensor_dims: tensor 'blk.1.ffn_down_exps.weight' has wrong shape; expected 2688, 4096, 512, got 2688, 1024, 512, 1
llama_model_load_from_file_impl: failed to load model
fish: Job 1, './koboldcpp-linux-x64' terminated by signal SIGSEGV (Address boundary error)
Do you know if this is a problem with the model or maybe koboldcpp?
Works perfect with llama.cpp b8356 with the following command on my HP ZB Ultra 14 G1a RAIM+P 128GB on Ubuntu 24.04.4 LTS.
Command:
llama-b8356-bin-ubuntu-vulkan-x64/llama-b8356/llama-server -hf bartowski/nvidia_Nemotron-3-Super-120B-A12B-GGUF:Q4_K_M -fa 1 --no-mmap