Runs great on the M2 96GB

#1
by jac-jim - opened

Running on battery, here are the speeds:

llama_perf_sampler_print:    sampling time =      71.98 ms /  1130 runs   (    0.06 ms per token, 15697.93 tokens per second)
llama_perf_context_print:        load time =   29767.26 ms
llama_perf_context_print: prompt eval time =     340.96 ms /    40 tokens (    8.52 ms per token,   117.32 tokens per second)
llama_perf_context_print:        eval time =   20637.74 ms /  1089 runs   (   18.95 ms per token,    52.77 tokens per second)
llama_perf_context_print:       total time =   55587.54 ms /  1129 tokens
llama_perf_context_print:    graphs reused =       1084
Interrupted by user

Wow! I can't wait to use this at work on Monday! :)

Sign up or log in to comment