Search Results for "readability"

...The repo is organized with a training pipeline (dataset preprocessing, model definition, optimizer, training loop) and inference script so you can train a small GPT on text datasets like Shakespeare or custom corpora. It emphasizes readability and clarity: the training loop is cleanly written, and the code avoids heavy abstractions, letting students follow the architecture step by step. While simple, it can still train non-trivial models on modern GPUs and generate coherent text. The project has become widely used in tutorials, courses, and experiments for people learning how transformers work under the hood.

A cross-platform text analysis program written in Python and Free Pascal/Lazarus which scans a whole text file (in plain text, HTML, EPUB, or ODT formats) and ranks all used words according to frequency, performing a quantitative analysis of the text using Shannon-Weaver information statistic and Zipf power law function. It counts words, sentences, chars, spaces, and syllables. Also computes readability indexes (Gunning-Fog, Coleman-Liau, Automated Readability Index (ARI), SMOG grade, Flesch–Kincaid grade level and Flesch Reading Ease).