ninjasaid13B to LocalLLaMA@poweruser.forumEnglish · 10 months agoLQ-LoRA: Low-rank Plus Quantized Matrix Decomposition for Efficient Language Model Finetuningarxiv.orgexternal-linkmessage-square2fedilinkarrow-up11arrow-down10
arrow-up11arrow-down1external-linkLQ-LoRA: Low-rank Plus Quantized Matrix Decomposition for Efficient Language Model Finetuningarxiv.orgninjasaid13B to LocalLLaMA@poweruser.forumEnglish · 10 months agomessage-square2fedilink
minus-squaremcmoose1900BlinkfedilinkEnglisharrow-up1·10 months agoAmazing! And they published the code. Also, the omniquant paper they linked is amazing! They hooked some super quantization into MLC, apparently?
Amazing! And they published the code.
Also, the omniquant paper they linked is amazing! They hooked some super quantization into MLC, apparently?