pankajmathur

/

orca_mini_13b

like 99

nah .... ggml ( llama.cpp ) version is far more better as can use much better precision like 2.3 4,5,6, or 8 bit and still works on any PC even without GPU than that ancient 4 bit gpto.

Thanks @flashvenom that was fast. Somehow the url gives 404 error.

I deleted mine since @TheBloke released models as well. No point having two copies to confuse folks