https://chkda.github.io/tags/llm/ Recent content in Llm on Chhaya's Space https://chkda.github.io/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E https://chkda.github.io/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E Hugo -- 0.168.0 en-us Sat, 24 Oct 2026 00:00:00 +0000 https://chkda.github.io/posts/lora_qlora/ Sat, 24 Oct 2026 00:00:00 +0000 https://chkda.github.io/posts/lora_qlora/ <h1 id="why-do-we-need-lora-and-qlora">Why do we need LoRA and QLoRA?</h1> <p>Say I have a pretrained language model and want to adapt it to my own dataset. Maybe I want it to answer questions in a particular domain or follow a specific response format. I can fine-tune it, but how much GPU memory will that take?</p> <p>With <strong>full fine-tuning</strong>, we update all the model&rsquo;s weights. Those weights are only part of the memory bill: training also needs gradients, optimizer states, and intermediate activations. A model that fits on my GPU for inference may still be too large to fine-tune on it.</p>