LoRA Speedrun – a public wall-clock leaderboard for fine-tuning techniques
Why CEREBRO kept it
Fine-tuning leaderboard; direct LLM optimization mechanics
The text below is an automated extraction of the article at https://github.com/Saivineeth147/lora-speedrun, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (github.com).
How fast can you LoRA-fine-tune Qwen2.5-1.5B to ≥ 57% on GSM8K — on a single L40S? This is modded-nanogpt for fine-tuning: a frozen task, frozen hardware, and a public leaderboard of wall-clock records. Every record is independently re-run 3× with fresh seeds on identical hardware before it counts. Attempting and verifying are free: official timing runs on a Modal L40S sandbox, and Modal's free monthly compute credits cover full runs — so anyone can compete, and anyone can re-verify any record with one command. Current record: 6m 05s by @Saivineeth147 — Sequence packing + completion-only loss
Backlinks
Appeared in 1 briefing