# Efficient Online Reinforcement Learning Fine-Tuning Need Not Retain Offline Data

arXiv: https://arxiv.org/abs/2412.07762 · HTML: https://mlsanity.com/paper/2412.07762/

## Key facts

- Cited by 63 papers in the map (data through 14 Oct 2026).
- Placed in Offline RL algorithms, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 52 of its 59 references that are in the map.
- First posted on arXiv 10 Dec 2024.

## Placed in

- Offline RL and imitation learning from data / [Offline RL algorithms](https://mlsanity.com/offlinerl/offline-rl-algorithms/)
- Deep reinforcement learning methods / [Deep RL algorithms and representations](https://mlsanity.com/deeprl/deep-rl-algorithms-and-representations/)

## Builds on (most cited first, 30 of 52)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Conservative Q-Learning for Offline Reinforcement Learning](https://mlsanity.com/paper/2006.04779/) (2020)
- [D4RL: Datasets for Deep Data-Driven Reinforcement Learning](https://mlsanity.com/paper/2004.07219/) (2020)
- [A Minimalist Approach to Offline Reinforcement Learning](https://mlsanity.com/paper/2106.06860/) (2021)
- [Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations](https://mlsanity.com/paper/1709.10087/) (2017)
- [AWAC: Accelerating Online Reinforcement Learning with Offline Datasets](https://mlsanity.com/paper/2006.09359/) (2020)
- [RL2: Fast Reinforcement Learning via Slow Reinforcement Learning](https://mlsanity.com/paper/1611.02779/) (2016)
- [Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables](https://mlsanity.com/paper/1903.08254/) (2019)
- [Relay Policy Learning: Solving Long-Horizon Tasks via Imitation and Reinforcement Learning](https://mlsanity.com/paper/1910.11956/) (2019)
- [Efficient Online Reinforcement Learning with Offline Data](https://mlsanity.com/paper/2302.02948/) (2023)
- [Randomized Ensembled Double Q-Learning: Learning Fast Without a Model](https://mlsanity.com/paper/2101.05982/) (2021)
- [Diffusion Policy Policy Optimization](https://mlsanity.com/paper/2409.00588/) (2024)
- [The Primacy Bias in Deep Reinforcement Learning](https://mlsanity.com/paper/2205.07802/) (2022)
- [Offline Reinforcement Learning with Fisher Divergence Critic Regularization](https://mlsanity.com/paper/2103.08050/) (2021)
- [Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning](https://mlsanity.com/paper/2303.05479/) (2023)
- [Offline-to-Online Reinforcement Learning via Balanced Replay and Pessimistic Q-Ensemble](https://mlsanity.com/paper/2107.00591/) (2021)
- [Precise and Dexterous Robotic Manipulation via Human-in-the-Loop Reinforcement Learning](https://mlsanity.com/paper/2410.21845/) (2024)
- [Hybrid RL: Using Both Offline and Online Data Can Make RL Efficient](https://mlsanity.com/paper/2210.06718/) (2022)
- [Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning](https://mlsanity.com/paper/2106.04895/) (2021)
- [SERL: A Software Suite for Sample-Efficient Robotic Reinforcement Learning](https://mlsanity.com/paper/2401.16013/) (2024)
- [Revisiting the Minimalist Approach to Offline Reinforcement Learning](https://mlsanity.com/paper/2305.09836/) (2023)
- [Continual World: A Robotic Benchmark For Continual Reinforcement Learning](https://mlsanity.com/paper/2105.10919/) (2021)

## Built on by (most cited first, 30 of 63)

- [ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policy](https://mlsanity.com/paper/2502.05450/) (2025)
- [Reinforcement Learning with Action Chunking](https://mlsanity.com/paper/2507.07969/) (2025)
- [Self-Improving Vision-Language-Action Models with Data Generation via Residual RL](https://mlsanity.com/paper/2511.00091/) (2025)
- [A Vision-Language-Action-Critic Model for Robotic Real-World Reinforcement Learning](https://mlsanity.com/paper/2509.15937/) (2025)
- [GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation](https://mlsanity.com/paper/2512.01801/) (2025)
- [Hyperspherical Normalization for Scalable Deep Reinforcement Learning](https://mlsanity.com/paper/2502.15280/) (2025)
- [Q-learning with Adjoint Matching](https://mlsanity.com/paper/2601.14234/) (2026)
- [Bigger, Regularized, Categorical: High-Capacity Value Functions are Efficient Multi-Task Learners](https://mlsanity.com/paper/2505.23150/) (2025)
- [Fast Adaptation with Behavioral Foundation Models](https://mlsanity.com/paper/2504.07896/) (2025)
- [Value Flows](https://mlsanity.com/paper/2510.07650/) (2025)
- [Intention-Conditioned Flow Occupancy Models](https://mlsanity.com/paper/2506.08902/) (2025)
- [Penalizing Infeasible Actions and Reward Scaling in Reinforcement Learning with Offline Data](https://mlsanity.com/paper/2507.08761/) (2025)
- [Simulation Distillation: Pretraining World Models in Simulation for Rapid Real-World Adaptation](https://mlsanity.com/paper/2603.15759/) (2026)
- [SimLauncher: Launching Sample-Efficient Real-world Robotic Reinforcement Learning via Simulation Pre-training](https://mlsanity.com/paper/2507.04452/) (2025)
- [The Three Regimes of Offline-to-Online Reinforcement Learning](https://mlsanity.com/paper/2510.01460/) (2025)
- [Fine-tuning Flow Matching Generative Models with Intermediate Feedback](https://mlsanity.com/paper/2510.18072/) (2025)
- [What Makes Value Learning Efficient in Residual Reinforcement Learning?](https://mlsanity.com/paper/2602.10539/) (2026)
- [Adaptive Action Chunking via Multi-Chunk Q Value Estimation](https://mlsanity.com/paper/2605.10044/) (2026)
- [Robust Policy Expansion for Offline-to-Online RL under Diverse Data Corruption](https://mlsanity.com/paper/2509.24748/) (2025)
- [Adaptive Replay Buffer for Offline-to-Online Reinforcement Learning](https://mlsanity.com/paper/2512.10510/) (2025)
- [Can We Really Learn One Representation to Optimize All Rewards?](https://mlsanity.com/paper/2602.11399/) (2026)
- [What Matters for Simulation to Online Reinforcement Learning on Real Robots](https://mlsanity.com/paper/2602.20220/) (2026)
- [When Life Gives You BC, Make Q-functions: Extracting Q-values from Behavior Cloning for On-Robot Reinforcement Learning](https://mlsanity.com/paper/2605.05172/) (2026)
- [QPILOTS: Efficient Test-Time Q-Steering for Flow Policies](https://mlsanity.com/paper/2606.14801/) (2026)
- [Rainbow-DemoRL: Combining Improvements in Demonstration-Augmented Reinforcement Learning](https://mlsanity.com/paper/2603.27400/) (2026)
- [WOMBET: World Model-Based Experience Transfer for Robust and Sample-efficient Reinforcement Learning](https://mlsanity.com/paper/2604.08958/) (2026)
- [RankQ: Offline-to-Online Reinforcement Learning via Self-Supervised Action Ranking](https://mlsanity.com/paper/2605.11151/) (2026)
- [Beyond Action Residuals: Real-World Robot Policy Steering via Bottleneck Latent Reinforcement Learning](https://mlsanity.com/paper/2605.19919/) (2026)
- [ReCoVLA: VLM-Guided Reward Compilation for Failure Recovery in Vision-Language-Action Policies](https://mlsanity.com/paper/2606.09630/) (2026)
- [Do You Really Need to Pretrain Q-Functions for Online RL Fine-Tuning?](https://mlsanity.com/paper/2607.27203/) (2026)

## Abstract (arXiv)

The modern paradigm in machine learning involves pre-training on diverse data, followed by task-specific fine-tuning. In reinforcement learning (RL), this translates to learning via offline RL on a diverse historical dataset, followed by rapid online RL fine-tuning using interaction data. Most RL fine-tuning methods require continued training on offline data for stability and performance. However, this is undesirable because training on diverse offline data is slow and expensive for large datasets, and in principle, also limit the performance improvement possible because of constraints or pessimism on offline data. In this paper, we show that retaining offline data is unnecessary as long as we use a properly-designed online RL approach for fine-tuning offline RL initializations. To build this approach, we start by analyzing the role of retaining offline data in online fine-tuning. We find that continued training on offline data is mostly useful for preventing a sudden divergence in the value function at the onset of fine-tuning, caused by a distribution mismatch between the offline data and online rollouts. This divergence typically results in unlearning and forgetting the benefits of offline pre-training. Our approach, Warm-start RL (WSRL), mitigates the catastrophic forgetting of pre-trained initializations using a very simple idea. WSRL employs a warmup phase that seeds the online RL run with a very small number of rollouts from the pre-trained policy to do fast online RL. The data collected during warmup helps ``recalibrate'' the offline Q-function to the online distribution, allowing us to completely discard offline data without destabilizing the online RL fine-tuning. We show that WSRL is able to fine-tune without retaining any offline data, and is able to learn faster and attains higher performance than existing algorithms irrespective of whether they retain offline data or not.