Full threadxfalcox·Is it just me or the 2 Llama 3.1 providers are the two know for running quantization levels that undermine Llama perf ?View on HN