• Taichi NeRF released: Try it now! It offers various applications and features.
• High-performance parallel programming in Python: Helps write portable and high-performance parallel programs.
• Easy to learn: Shares Python syntax, with a single decorator turning functions into optimized code.
• High-performance: JIT compiler compiles Python to fast machine code for parallel execution.
• Integrated into Python ecosystem: Seamlessly interoperates with NumPy, PyTorch, etc.
• Universal deployment: AOT mode enables deployment on platforms without Python.
• Key features: One-billion-particle MPM simulation, automatic differentiation, flexible SNode system, and spatially sparse data structures.
• Used by organizations: ETH Zürich in Physically-based Simulation course, University of Utah in a paper, OPPO for mobile wallpapers, and Kuaishou as high-performance mobile GPU solution.