Revision History
QLoRA · 9 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: Corrected a memory table falsely attributed to the QLoRA paper, wrong adapter and optimizer memory figures for the 65B model, the paper's learning-rate schedule (constant, not cosine), a misreported Hugging Face T4 benchmark, GaLore's 24 GB result (pre-training, not fine-tuning), Tim Dettmers' affiliations, and an unsupported claim that Zephyr and other named community models were QLoRA-trained (plus 24 minor fixes)
Version 9 · Nov 2, 2026, 11:03 AM
- Improve internal links (2026-10-13)
Version 8 · Oct 13, 2026, 10:18 AM
- Added 6 contextual internal links
Version 7 · Jul 23, 2026, 03:28 PM
- Pre-update snapshot before enhancement to 7786 words.
Version 3 · May 18, 2026, 12:27 AM