Awesome GitHub RepositoriesGreedy Decoding Strategies

This project is an educational platform and research toolkit designed to teach deep learning through a combination of mathematical theory, visual diagrams, and executable code. It provides a comprehensive environment for building, training, and evaluating neural networks, grounding complex concepts in interactive computational notebooks that allow for hands-on experimentation. The framework distinguishes itself by interleaving theoretical foundations—including linear algebra, calculus, and probability—with practical implementations across multiple industry-standard libraries. It supports flex

هذا إطار عمل ومكتبة نماذج للتشفير وفك التشفير (encoder-decoder) مبنية على TensorFlow تُستخدم لتعيين تسلسلات الإدخال إلى تسلسلات الإخراج. يعمل كمُعيّن تسلسلات للتعلم العميق مصمم لتحويل البيانات التسلسلية من مجال إلى آخر. توفر المكتبة أدوات لتنفيذ نمذجة التسلسل إلى تسلسل عبر مجالات متعددة، بما في ذلك الترجمة الآلية العصبية، وتلخيص النصوص التلقائي، وتوليد تعليقات توضيحية للصور. يدمج إطار العمل الشبكات العصبية المتكررة ويستخدم السياق القائم على الانتباه (attention-based) لوزن تسلسلات الإدخال. يدعم استراتيجيات فك تشفير متعددة، بما في ذلك بحث الشعاع (beam search) وفك التشفير الجشع (greedy decoding)، مع تنفيذ العمليات الرياضية عبر حسابات الرسم البياني في TensorFlow.

This project is a deep learning poetry generator designed to create traditional Chinese couplets. It utilizes a sequence-to-sequence neural network architecture to map input text sequences to matching output sequences, functioning as a text generation model and an inference web service. The system features a neural text ranking mechanism that evaluates candidate outputs based on length consistency and character patterns to ensure structural alignment. It also includes a content filtering process that scans generated text against forbidden word lists to remove sensitive or inappropriate materi

Neuraltalk is an automated image captioning system that generates natural language descriptions for images. It utilizes a deep learning model that integrates a pretrained convolutional neural network for visual feature extraction with a recurrent neural network decoder to produce text sequences. The project provides a full workflow for training and evaluating captioning models, including weight optimization via backpropagation and gradient descent. It includes tools for measuring caption accuracy by comparing generated text against reference descriptions. The system covers data preprocessing

CTranslate2 is a C++ inference engine and runtime for Transformer models, designed to execute models on both CPU and GPU with optimizations for speed and memory efficiency. It functions as a model format converter, quantization tool, and REST API server, enabling deployment of neural machine translation, automatic speech recognition, and text generation models. The engine distinguishes itself through a suite of runtime optimizations including layer fusion, weight-matrix quantization, batch-by-length grouping, and a caching allocator that reuses GPU memory. It supports tensor-parallel model di

picoGPT is a lightweight, low-level runtime environment and inference engine designed to load pre-trained checkpoints and execute generative transformer model inference. It provides a minimal implementation of the generative pre-trained transformer architecture to facilitate local language model execution. The project includes a C++ machine learning library for converting model parameters and executing greedy token generation without heavy external dependencies. It handles remote asset synchronization by downloading pre-trained weights, hyperparameters, and vocabulary files from remote server