CrystalFlow

CrystalFlow is a flow-based generative model for crystalline materials using Continuous Normalizing Flows (CNFs) within the Conditional Flow Matching (CFM) framework. Published in Nature Communications 2025.

7. STRUCTURE PREDICTION 7.5 Generative Models VERIFIED 1 paper
Back to Mind Map Official Website

Overview

CrystalFlow is a flow-based generative model for crystalline materials using Continuous Normalizing Flows (CNFs) within the Conditional Flow Matching (CFM) framework. Published in Nature Communications 2025.

Reference Papers (1)

Full Documentation

Overview

CrystalFlow is a flow-based generative model for crystalline materials using Continuous Normalizing Flows (CNFs) within the Conditional Flow Matching (CFM) framework. Published in Nature Communications 2025.

Theoretical Basis

  • Continuous Normalizing Flows (CNFs)
  • Conditional Flow Matching (CFM)
  • Periodic structure handling
  • Lattice and coordinate generation
  • Property-guided generation

Key Capabilities

  • Crystal structure generation
  • Flow-based sampling
  • Property optimization
  • Stable structure generation
  • Nature Communications publication

Sources: Nature Communications (2025), arXiv:2412.11693

Key Strengths

Methodology:

  • Flow-based (not diffusion)
  • CNF framework
  • Efficient sampling

Performance:

  • Competitive results
  • Multiple benchmarks
  • Good stability

Publication:

  • Nature Communications 2025
  • Peer-reviewed
  • Well-documented

Inputs & Outputs

  • Input formats: Training structures, composition (generation)
  • Output data types: Generated crystal structures

Interfaces & Ecosystem

  • Framework: PyTorch
  • Datasets: Standard benchmarks
  • Evaluation: DFT validation

Workflow and Usage

  1. Train on crystal dataset
  2. Sample from flow model
  3. Generate structures
  4. Evaluate stability
  5. Validate with DFT

Performance Characteristics

  • GPU-accelerated
  • Efficient flow sampling
  • Good generation quality

Computational Cost

  • Training: GPU hours
  • Sampling: fast
  • Validation: DFT

Best Practices

  • Use appropriate training data
  • Generate multiple samples
  • Validate predictions
  • Check structural validity

Limitations & Known Constraints

  • Training data dependent
  • Recent development
  • Requires validation

Application Areas

  • Crystal structure generation
  • Materials discovery
  • Property-guided design
  • Generative materials science

Comparison with Other Codes

  • vs FlowMM: Similar approach, different implementation
  • vs DiffCSP: Flow vs diffusion
  • vs CDVAE: Different generative framework
  • Unique strength: CNF/CFM framework, Nature Comms 2025

Community and Support

  • Open-source (GitHub)
  • Academic development
  • Recent publication

Verification & Sources

Primary sources:

  1. GitHub: https://github.com/ixsluo/CrystalFlow
  2. Publication: Nature Communications (2025)
  3. arXiv: https://arxiv.org/abs/2412.11693

Confidence: VERIFIED

Verification status: ✅ VERIFIED

  • Website: ACTIVE (GitHub)
  • Source: OPEN
  • Development: RECENT (2025)
  • Applications: Crystal structure generation

Related Tools in 7.5 Generative Models