3 ms·
Building the New Hardware for AI Model Inference [video]
- qrios 3y agoInteresting approach to load the model direct. But no explanation on the video or on the company page[1] about the processing units itself. The numbers from the page (2.3x faster, 1/2 the power) are plausible for top notch FPGAs. [1] https://www.positron.ai/product https://www.positron.ai/product