Search Results for "gt-main"

Python library for comparing the distance between two or more sequences by many algorithms. For main algorithms, text distance try to call known external libraries (fastest first) if available (installed in your system) and possible (this implementation can compare this type of sequences). Install text distance with extras for this feature. Textdistance use benchmark results for algorithm optimization and try to call the fastest external lib first (if possible).

DualPipe is a bidirectional pipeline parallelism algorithm open-sourced by DeepSeek, introduced in their DeepSeek-V3 technical framework. The main goal of DualPipe is to maximize overlap between computation and communication phases during distributed training, thus reducing idle GPU time (i.e. “pipeline bubbles”) and improving cluster efficiency. Traditional pipeline parallelism methods (e.g. 1F1B or staggered pipelining) leave gaps because forward and backward phases can’t fully overlap with communication. ...

This collection of Python3 modules provides a large range of implemented decision aiding algorithms useful in the field of outranking digraphs based Multiple Criteria Decision Aid (MCDA), especially best choice, linear ranking and absolute or relative rating algorithms with multiple incommensurable criteria. Technical documentation and tutorials are available under the following link: https://digraph3.readthedocs.io/en/latest/ The tutorials introduce the main objects like digraphs, outranking digraphs and performance tableaux. There is also a tutorial provided on undirected graphs. Some tutorials are problem oriented and show how to compute the winner of an election, how to build a best choice recommendation, or how to linearly rank or rate with multiple incommensurable performance criteria. ...

...The project demonstrates how reinforcement learning can be used to automatically discover efficient algorithms for matrix multiplication — a fundamental operation in computer science and numerical computation. The repository is organized into four main components: algorithms, benchmarking, nonequivalence, and recombination. These contain implementations of the discovered matrix multiplication algorithms, tools to benchmark their real-world performance, proofs of nonequivalence among thousands of solutions, and methods for decomposing larger problems into smaller factorizations. ...

Unlike the other two, openai/baselines is not currently a maintained or prominent repo in the OpenAI organization (and I found no strong reference in OpenAI’s main GitHub). Historically, “baselines” repositories are often used for baseline implementations of reinforcement learning algorithms or reference models (e.g. in the RL domain). If there was an OpenAI “baselines” repo, it might have contained reference implementations for reinforcement learning or model policy baselines to compare new work against. ...

...Coach collects statistics from the training process and supports advanced visualization techniques for debugging the agent being trained. Coach supports many state-of-the-art reinforcement learning algorithms, which are separated into three main classes - value optimization, policy optimization, and imitation learning. Coach supports a large number of environments which can be solved using reinforcement learning.

...The system allows researchers to study how models can improve labeling efficiency by selectively querying the most informative data points rather than relying on uniformly sampled training sets. The main experiment runner (run_experiment.py) supports a wide range of configurations, including batch sizes, dataset subsets, model selection, and data preprocessing options. It includes several established active learning strategies such as uncertainty sampling, k-center greedy selection, and bandit-based methods, while also allowing for custom algorithm implementations. ...

LZO is a compression library which is widely used around the world. The main problem with LZO is that it is absolutely not human readable. People have done crazy stuff to get LZO to run in their language. Usually it implies inline assembly or trying to execute data which actually contains machine code. This is sick. Whoever is responsible for this sorry situation ought to be ashamed. So I'm going to deobfuscate LZO and provide a ISC implementation of this algorithm in Python and C. ...