Users working on PyTorch will be able to use ROCm 5.7 on Ubuntu Linux and take advantage of the power of the Radeon RX 7900 XTX and Radeon PRO W7900 GPUs.
In an effort to make intelligent development more accessible, Advanced Micro Devices (AMD) is introducing new features.
After releasing ROCm 5.7, a new version of its open software ecosystem for accelerated computing, it extends support for PyTorch machine learning on graphics cards that are based on AMD RDNA 3.
Those working with models or algorithms in PyTorch will be able to use ROCm 5.7 on Ubuntu Linux and take advantage of the parallel computing power of the Radeon RX 7900 XTX and Radeon PRO W7900 GPUs. These GPUs offer 192 dedicated acceleration units, delivering double the performance per compute unit over the previous generation.
In doing so, AMD enables machine learning from the desktop for training and inference.