[2305.14314] QLoRA: Efficient Finetuning of Quantized LLMs
7
Public whispers
6
Contributors
2026-07-22 10:05:02
First whispered
Public whispers on this page
Text Highlight2026-08-11 18:38:31
Original Highlight Excerpt
"QLoRA backprop"
Whisper Note
My 48GB card finally feels useful for something other than gaming.
Text Highlight2026-07-22 16:29:02
Original Highlight Excerpt
"arXivLabs is a framework that allows collaborators to develop and share new arXiv features"
Whisper Note
Sounds like a cool way for the community to pitch in on features.
Text Highlight2026-07-22 13:26:02
Original Highlight Excerpt
"preserving full 16-bit finetuning task performance"
Whisper Note
I just hope it's not another paper that overpromises on benchmarks.
Text Highlight2026-07-22 13:17:02
Original Highlight Excerpt
"preserving full 16-bit finetuning task performance"
Whisper Note
Sounds too good to be true, but if it works, that's a game changer.
Text Highlight2026-07-22 10:23:02
Original Highlight Excerpt
"reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU"
Whisper Note
Tried it on a 13B model, memory drop was insane. Game changer for my lab.
Text Highlight2026-07-22 10:14:02
Original Highlight Excerpt
"reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU"
Whisper Note
But does it really keep full 16-bit performance? I'd like to see more tests.
Text Highlight2026-07-22 10:05:02
Original Highlight Excerpt
"reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU"
Whisper Note
Finally can train big models without selling a kidney for GPUs.
Share this page's whispers
Short link
https://domwhisper.com/s/2e95ee3ff8d6Embed snippet
<iframe src="https://domwhisper.com/embed/2e95ee3ff8d6" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>See what people are discussing on arxiv.org
Install DomWhisper to view live whispers as you browse, and join the discussion.
Get the extension