unsloth/DeepSeek-R1-0528-GGUF · Problem with llama-server - llama_model_load: error loading model: check_tensor_dims: tensor 'blk.0.attn_q

I just compiled the latest llama-server from the official repository and I am trying to run DeepSeek-R1-0528-UD-Q2_K_XL.

The command line is:
llama-server.exe -m "D:\AI_Projects\DeepSeek-R1-0528-GGUF\DeepSeek-R1-0528-UD-Q2_K_XL-00001-of-00006.gguf" --temp 0.6 --top-p 0.95 -ngl 50

I get the error:
llama_model_load: error loading model: check_tensor_dims: tensor 'blk.0.attn_q_b.weight' has wrong shape; expected 1536, 73728, got 1536, 24576, 1, 1

The workstation has 2 x 24GB GPU and 256GB RAM, I assume it is not a memory problem.

Is llama-server working for you?

unsloth
/

DeepSeek-R1-0528-GGUF

Problem with llama-server - llama_model_load: error loading model: check_tensor_dims: tensor 'blk.0.attn_q_b.weight' has wrong shape