Interesting to see an #implementation of #mini #gpt https://github.com/karpathy/minGPT Could lead to some cool #applications that don't require a #supercomputer to #train and #run !
Pretty cool that #yandex have #opensourced a 100 #billion #parameter #model https://github.com/yandex/YaLM-100B That said, you need a #supercomputer to actually run the thing.