AI · Official
Why Fine-Tuning Jobs Run Longer Than FLOP Estimates Predict on Shared GPUs
Hugging Face blog ·

At a glance
A team tested how well the usual FLOP-based estimate predicts how long a fine-tuning job takes on shared GPUs. Using 272 real runs on H200 GPUs, the simple estimate explained little, while a learned model using pre-dispatch information did far better. The gap comes from overheads like setup, evaluations, and checkpoint saves that the formula misses.
Key points
- The team studied 272 real supervised fine-tuning jobs on NVIDIA H200 GPUs, totaling about 7,382 GPU-hours.
- The common FLOP-based estimate explained little of the observed run times (R² of 0.31, mean error of 12.41 hours).
- A CatBoost model using only information available before dispatch reached R² of 0.98 with a mean error of 1.71 hours.
- The formula misses overheads such as setup, periodic evaluations, checkpoint uploads, and I/O competition on shared clusters.
Why it matters
Better time estimates help teams schedule scarce GPUs with less idle time and fewer delays for queued jobs.
Original source
Why your fine-tuning job takes longer than the FLOPs say
Hugging Face blog · Nov 19, 03:53 KST
Read on Hugging Face blog ↗AI summarized the original article. Check the original for accuracy. Copyright belongs to Hugging Face blog.
More from this field
- Stretto Builds AI Voice Agent Tara on Amazon Nova 2 Sonic for Bankruptcy Intake
Stretto, a company that runs the Bankruptcy.com site, built a phone-based AI voice agent named Tara. Tara talks with callers who are considering bankruptcy and collects detailed financial information in conversation. The agent is built on Amazon Nova 2 Sonic, a speech model that handles voice input and output in one system.
Amazon AWS Machine Learning blog · Nov 19, 03:58 KST - AWS Quick User Group DC Meetup Shows How Companies Prepare Data for AI Agents
AWS held an inaugural AWS Quick User Group meetup in Arlington, Virginia, on October 7, 2026. Data leaders and developers talked about how to prepare company data so AI agents can use it. One firm, Financial Independence Group (FIG), shared how it is using AWS Quick for its analytics.
Amazon AWS Machine Learning blog · Nov 19, 03:54 KST - Microsoft Launches Seven Cheaper In-House AI Models to Replace Outside Ones
Microsoft AI has released seven new in-house models for coding, security, images, speech, and routing, all at once at its Ignite conference. The company says they match rival models at a much lower cost, and it is aiming them at AI agents that run around the clock.
The Decoder · Nov 19, 03:47 KST