build: 5080 (78a1ba0a) with Apple clang version 16.0.0 (clang-1600.0.26.6) for arm64-apple-darwin24.2.0
llama_model_load_from_file_impl: using device Metal (Apple M3 Pro) - 27647 MiB free
llama_model_loader: loaded meta data with 33 key-value pairs and 292 tensors from ./DeepSeek-R1-Distill-Llama-8B-Q3_K_L.gguf (version GGUF V3 (latest))

Final Winogrande score(750 tasks): 67.7333 +/- 1.7082

llama_perf_context_print:        load time =     299.54 ms
llama_perf_context_print: prompt eval time =   75587.39 ms / 22192 tokens (    3.41 ms per token,   293.59 tokens per second)
llama_perf_context_print:        eval time =       0.00 ms /     1 runs   (    0.00 ms per token,      inf tokens per second)
llama_perf_context_print:       total time =   76080.69 ms / 22193 tokens
ggml_metal_free: deallocating