Instructions for llm-cli obsolete?

[MSSRPRAD]

I get this error when I do:

llm gptneox infer --model-path '/home/pradyumnamalladi/Downloads/RedPajama-INCITE-Base-3B-v1-q4_0.bin' -p "Hello!";
⣯ Loaded hyperparametersError: 
   0: Could not load model
   1: non-specific I/O error
   2: failed to fill whole buffer

Backtrace omitted. Run with RUST_BACKTRACE=1 environment variable to display it.
Run with RUST_BACKTRACE=full to include source snippets.

Pls guide on how to run these models now:

RedPajama-INCITE-Base-3B-v1-q4_0.bin
llama-2-7b-chat.ggmlv3.q2_K.bin
llama-7b.ggmlv3.q2_K.bin

[philpax]

Hi there, thanks for reporting this! Yes, the released CLI is quite out of date now; I've updated the README to indicate that you should install from source instead. Does that solve the problem for you?

[MSSRPRAD]

Yeah it worked great when I cloned the repo and ran it so it's fine.

Does this run as fast as llama.cpp btw?