mixtral inference numbers
see the others: - [llama inference numbers] - [llama2 inference numbers] i am not running perplexity scores on these (that's more important for a quantization comparison chart, rather than performance parameter tweaking), just want to find out how …