AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

FastGRPO: Accelerating Policy Optimization via Concurrency-aware Speculative Decoding and Online Draft Learning

Yizhou Zhang, Ning Lv, Teng Wang, Jisheng Dang

ICLR 2026 · Reinforcement learning › RL from feedback · approach: GRPO stability and variants

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add online draft learning (ablation)FastGRPO · GRPO on 3 math datasets (average)end-to-end speedup: 1.74 -> 2.49Improved
add concurrency-aware drafting and verification (ablation)FastGRPO · GRPO on 3 math datasets (average)end-to-end speedup: 2.21 -> 2.49Improved
speculative decoding with concurrency-aware settings versus vanilla with early termination (ablation)GRPO generation · GRPO on 3 math datasets (average)generation speedup: 1.51 -> 2.78Improved
fixed small-batch speculative config used at larger batch sizes (ablation)speculative decoding · batch size scalingspeedup: speedup drops below 1.0x at higher concurrencyHurt
online draft learning does not harm general-purpose tasks (ablation)draft model · MT-bench and HumanEvalspeedup ratio: 1.71 -> 1.72 (MT-bench); 2.01 -> 2.05 (HumanEval)No clear effect
online draft learning on SimpleRL-Abel data (ablation)draft model · GSM8K (in-domain)average acceptance length: 3.82 -> 4.55Improved
skip draft pretraining, train draft only online (ablation)FastGRPO · SimpleRL-Abel-Level3to5, Qwen2.5-7B-Instructaverage accepted lengthNo clear effect
apply FastGRPO to GRPO variants DAPO and GPGFastGRPO · 3 math datasets, Qwen2.5-7B-Instructend-to-end speedup: 2.34 (DAPO), 2.57 (GPG)Improved
use off-the-shelf speculative decoding (EAGLE-2, HASS, EAGLE-3) in GRPOGRPO generation · 3 math datasetsend-to-end speedup: average E2E 1.12 to 1.27No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy