NVIDIA NIM for Boltz-2

Boltz-2

• Boltz-2 NIM Overview
  • Advantages of NIMs
• Release Notes
  • Release 1.9.0
    • Summary
    • Key Features
    • Removed Configuration
    • Notes
  • Release 1.8.0
    • Summary
    • Key Features
    • Removed Configuration
    • Notes
  • Release 1.7.0
    • Summary
    • Key Features
  • Release 1.6.0
    • Summary
    • Key Features
  • Release 1.5.0
    • Summary
    • Key Features
  • Release 1.4.0
    • Summary
    • Key Features
  • Release 1.3.0
    • Summary
    • Key Features
  • Release 1.2.0
    • Summary
    • Key Features
    • Notes and Limitations
  • Release 1.1.0
    • Summary
    • Notes and Limitations
• Support Matrix
  • Supported Hardware
    • Minimum System Hardware Requirements
    • Supported NVIDIA GPUs
    • Model Profile
  • Testing Locally Available Hardware
• Getting Started
  • Prerequisites
    • Installing curl, jq, and the Python requests module
    • NGC Authentication
      • Generate an API key
      • Export the API key
      • Docker Login to NGC
  • Starting the NIM Container
  • Runtime Parameters for the Container
  • Caching Models
  • Stopping the Container
  • Running on Slurm With Enroot
    • Environment
    • 1. Check Enroot is Available
    • 2. Create and Add NGC Credentials
    • 3. Configure and Run the Slurm Job Script
    • 4. Call the API From Your Machine
  • Next Steps
• Inferencing with the NIM
  • Before You Start
  • NIM API Reference
    • OpenAPI Specification
    • Example Notebooks
    • Predict a Protein Structure
      • Input Parameters
      • Outputs
      • Validation and Errors
      • Examples
        • Basic Protein Structure Prediction
        • Protein-Ligand Complex Prediction
      • Example with Multiple Features
        • PAE and PDE Outputs
    • Check Readiness
      • Input parameters
      • Outputs
      • Example
• Configuring the Boltz-2 NIM
  • GPU Selection
  • Environment Variables
  • NIM Telemetry
    • NIM_TELEMETRY_MODE
    • NIM_TELEMETRY_ENABLE_ON_RTX
  • Volumes
  • Logging Configuration
    • Example: Running with Logging Configuration
    • Example: Debug Logging for Troubleshooting
• Loading Custom / Fine-Tuned Boltz-2 Checkpoints
  • How It Works
  • Quickstart
  • Renaming Checkpoints
  • Notes and limitations
• Optimization with Boltz2-NIM
  • Automatic Profile Selection
    • Selecting a Profile Manually
    • Optimized Backend
  • Deploying the NIM on a Multi-GPU System
  • Adjusting Start-Time NIM Input Limits
    • Environment Variables
      • NIM_MAX_POLYMER_INPUTS
      • NIM_MAX_LIGAND_INPUTS
      • NIM_MAX_POLYMER_LENGTH
      • NIM_MAX_TEMPLATES_PER_POLYMER
      • NIM_MAX_MSA_SEQS
    • Usage
  • Optimization Parameters
    • Recycling Steps
    • Sampling Steps
    • Diffusion Samples
    • Max Parallel Samples
    • Full PAE and PDE Matrices
    • Step Scale
    • Binding Affinity Prediction Parameters
      • Affinity Sampling Steps
      • Affinity Diffusion Samples
      • Molecular Weight Correction
    • Ligand-Specific Performance Considerations
  • Querying the NIM Repeatedly
    • Running Repeated Queries Against the NIM Serially
• Performance in Boltz-2 NIM
  • NIM Accuracy
  • Factors Affecting NIM Performance
    • Hardware Factors
    • Input Complexity
    • Model Parameters
  • Performance Characteristics
    • Typical Runtimes
  • Performance Metrics (Boltz2 v1.6.0)
    • Configuration
    • Table 1: Performance Across the Supported NVIDIA Hardware Units
    • Table 2: Performance Across Optimization Backends
    • Table 3: Performance Impact From Structural Templates
  • Boltz2 v1.6.0 Performance Results on H100
    • Typical Runtimes
    • Structure Prediction Performance
    • Binding Affinity Prediction Performance
    • Performance Analysis
  • Performance Testing
    • Performance Optimization Tips
  • Troubleshooting Performance Issues
    • Common Issues and Solutions
      • General Performance Issues
      • Binding Affinity Specific Issues
• Deploy Boltz2 with Helm
  • Prerequisites
    • Required Components
    • Download the Helm Chart
  • Configure Helm
    • Core Configuration Options
  • Quick Start Deployment
    • 1. Create NGC Secrets
    • 2. Create Values File
    • 3. Install the Chart
    • 4. Verify Deployment
  • Storage
    • Persistent Volume Claims (Recommended)
    • Alternative Storage Options
      • NFS Storage
      • HostPath (development)
  • GPU Configuration
    • Single GPU Deployment
    • GPU Selection
  • Service Configuration
    • Basic Service
    • Port Forwarding (Testing)
    • Ingress Configuration
  • Advanced Configuration
    • Autoscaling
    • Custom Probes
    • Security Context
  • Run Inference
  • Troubleshooting
    • Pod Stuck in Pending
    • Pod Fails to Start
    • Storage Issues
    • GPU Access
  • Uninstall
  • Additional Resources
• EULA