Tim Dettmers
Creator and maintainer of bitsandbytes; Assistant Professor at CMU and Research Scientist at Ai2
- Assistant Professor, Carnegie Mellon University12
- Research Scientist, Ai2 (Allen Institute for AI)13
Tim Dettmers works on making AI models practical to train and use with modest compute. He created and maintains bitsandbytes, a PyTorch library for k-bit quantization that provides 8-bit optimizers, LLM.int8() inference, and support for QLoRA fine-tuning. He is first author of the QLoRA paper, which fine-tunes a frozen 4-bit quantized model through low-rank adapters and reports fine-tuning a 65B-parameter model on a single 48GB GPU. He is an Assistant Professor at Carnegie Mellon University and a Research Scientist at Ai2.1452
Work
bitsandbytes (external site: github.com)
Created the library and maintains it; its README cites his papers for the 8-bit optimizers, LLM.int8(), and QLoRA features.124
QLoRA (external site: arxiv.org)
First author of the paper introducing QLoRA, a method for fine-tuning 4-bit quantized language models through low-rank adapters.5
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.