速度---1.81 tokens per second:
llama_perf_sampler_print: sampling time = 165.87 ms / 2240 runs ( 0.07 ms per token, 13504.47 tokens per second)
llama_perf_context_print: load time = 1642.30 ms
llama_perf_context_print: prompt eval time = 181728.90 ms / 878 tokens ( 206.98 ms per token, 4.83 tokens per second)
llama_perf_context_print: eval time = 753774.45 ms / 1361 runs ( 553.84 ms per token, 1.81 tokens per second)
llama_perf_context_print: total time = 1043833.55 ms / 2239 tokens