zhaoyl18 / SEIKO
SEIKO is a novel reinforcement learning method to efficiently fine-tune diffusion models in an online setting. Our methods outperform all baselines (PPO, classifier-based guidance, direct reward backpropagation) for fine-tuning Stable Diffusion.
☆18Updated 6 months ago
Alternatives and similar repositories for SEIKO:
Users that are interested in SEIKO are comparing it to the libraries listed below
- Code for the tutorial/review paper for RL-based-fine-tuniing. In this code, we especially focus on the design of biological sequences li…☆103Updated 4 months ago
- Official codebase for Exact Energy-Guided Diffusion Sampling via Contrastive Energy Prediction☆25Updated last year
- Official Code for Local Search GFlowNets (ICLR 2024 Spotlight)☆16Updated 5 months ago
- ☆62Updated last month
- ☆23Updated 10 months ago
- Derivative-Free Guidance in Diffusion Models with Soft Value-Based Decoding. For controlled generation in DNA, RNA, proteins, molecules (…☆26Updated 3 months ago
- Official codebase for Exact Energy-Guided Diffusion Sampling via Contrastive Energy Prediction (ICML 2023)☆45Updated last year
- ☆26Updated last year
- Educational implementation of the Discrete Flow Matching paper☆74Updated 5 months ago
- ☆31Updated 7 months ago
- Code for the paper Iterated Denoising Energy Matching for Sampling from Boltzmann Densities.☆46Updated this week
- ☆21Updated 6 months ago
- ☆28Updated last month
- ☆15Updated 9 months ago
- This repository is the official implementation of Bidirectional Learning for Offline Infinite-width Model-based Optimization (NeurIPS 202…☆15Updated 2 years ago
- Code for Paper "Think While You Generate: Discrete Diffusion with Planned Denoising"☆37Updated 2 months ago
- Graph Diffusion Policy Optimization☆30Updated 10 months ago
- Code for paper: "Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design"☆32Updated 3 months ago
- Implementation of Particle Guidance: non-I.I.D. Diverse Sampling with Diffusion Models☆65Updated last year
- Code for "Optimizing DDPM Sampling with Shortcut Fine-Tuning" (https://arxiv.org/abs/2301.13362), ICML 2023☆27Updated last year
- ☆12Updated 3 months ago
- ☆49Updated 7 months ago
- Official implementation of ICML'24 paper "Offline Multi-Objective Optimization".☆17Updated 3 months ago
- Code for the paper https://arxiv.org/abs/2402.04997☆65Updated 11 months ago
- Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold☆45Updated 4 months ago
- Simple Guidance Mechanisms for Discrete Diffusion Models☆19Updated last month
- ☆48Updated 4 months ago
- ☆15Updated last year
- Baselines for Model-Based Optimization installation fixes and compatible with newer AMPERE+ GPUs (e.g. 3090)☆11Updated last year
- NF-Layers for constructing neural functionals.☆81Updated last year