Hyperparameter optimization framework Optuna™ is an open-source automatic hyperparameter optimization framework which automates the trial-and-error process of optimizing the hyperparameters. It automatically finds optimal hyperparameter values based on an optimization target. Optuna is framework agnostic and can be used with most Python frameworks, including Chainer, Scikit-learn, Pytorch, etc. Optuna is used in PFN projects such as Kachaka autonomous mobile robot, Matlantis universal atomistic simulator and more.
MN-Core 2, a proprietary AI chip developed by PFN. MN-Core 2 features a higher local memory capacity per compute unit and greater flexibility in software control compared to GPUs, enabling significant speed improvements for simulation workloads for specific structures.
CATEGORY
Library for general-purpose GPU matrix calculation CuPy™ is an open-source matrix library accelerated with NVIDIA CUDA. It has a NumPy-compatible interface and uses CUDA-related libraries including cuBLAS, cuDNN, cuRand, cuSolver, cuSPARSE, cuFFT, and NCCL to make full use of the GPU architecture. CuPy was born as a GPU backend of Chainer, and it has been separated from Chainer into an independent package from 2017.
CATEGORY
