Cluster 440326

Libraries and models for predicting time series data using transformer-based deep learning architectures. The cluster centers on Chronos, a family of pretrained transformer models (tiny through large variants) designed for zero-shot time series forecasting across diverse datasets and domains. Resources here cover the practical application of modern NLP-inspired architectures to temporal prediction problems, including both the models themselves and supporting infrastructure for evaluation and deployment.

Models and tools for forecasting temporal sequences using transformer-based and machine learning approaches. The cluster centers on foundation models like Moirai and Granite TimeSeries that apply large-scale pre-trained architectures to time-series prediction tasks, alongside supporting utilities and datasets. Resources here span model implementations, evaluation frameworks, and applications across different scales and configurations.

Python libraries and machine learning frameworks for predicting temporal sequences using deep learning models. This cluster covers end-to-end time series analysis—from foundational forecasting architectures (neural networks, attention mechanisms, probabilistic methods) to applied forecasting systems and datasets. Repositories include general-purpose forecasting libraries like DARTS, specialized probabilistic models for clinical predictions, and research implementations exploring neural architectures for sequential data.