Reference-grade guide to shrinking and speeding up LLMs without retraining from scratch — numeric formats (FP8/INT8/INT4), PTQ methods (GPTQ, AWQ, SmoothQuant, bitsandbytes NF4, GGUF k-quants), KV-cache quantization, speculative decoding (Medusa/EAGLE/n-gram), and distillation — with concrete number
Reference-grade guide to shrinking and speeding up LLMs without retraining from scratch — numeric formats (FP8/INT8/INT4), PTQ methods (GPTQ, AWQ, SmoothQuant, bitsandbytes NF4, GGUF k-quants), KV-cache quantization, speculative decoding (Medusa/EAGLE/n-gram), and distillation — with concrete numbers, when each fits, and the quality cliffs.
Esta entrada no publica ningún paquete de npm, así que Forge no tiene un árbol de dependencias para ella. Es una carencia de cobertura, no una afirmación de que no tenga dependencias.