3 ms·
Fabrice Bellard: Fast, Standalone CPU Inference for Transformer Language Models
- h2odragon 5y agoalso see https://bellard.org/nncp/ https://bellard.org/nncp/ Lossless Data Compression with Neural Networks; which appears to be the immediate prompt.