One-Embedding-Fits-All: Selecting the Best Time-Series Forecasting Model from a Model Zoo

The paper introduces ZooCast, a framework that builds a model zoo of time‑series foundation models and uses a One‑Embedding‑Fits‑All paradigm to embed models and tasks into a unified space, enabling efficient zero‑shot selection that outperforms single models and full‑model ensembles on the GIFT‑Eval benchmark while remaining computationally lightweight.

GIFT-EvalOne-Embedding-Fits-AllTSFM

0 likes · 10 min read

One-Embedding-Fits-All: Selecting the Best Time-Series Forecasting Model from a Model Zoo

Alibaba Cloud Developer

Oct 22, 2020 · Artificial Intelligence

EasyTransfer: Alibaba’s Open‑Source Framework Boosting NLP Transfer Learning

Alibaba Cloud open‑sources EasyTransfer, a high‑performance deep transfer‑learning framework for NLP that unifies pre‑training, knowledge distillation, meta‑learning and distributed deployment, offering a rich ModelZoo, AppZoo and seamless integration with the PAI ecosystem.

Distributed TrainingEasyTransferNLP

0 likes · 15 min read

EasyTransfer: Alibaba’s Open‑Source Framework Boosting NLP Transfer Learning

AntTech

Jul 10, 2020 · Artificial Intelligence

SQLFlow: End‑to‑End AI Workflow Construction Using SQL

SQLFlow is an open‑source system that lets users describe and execute end‑to‑end AI tasks—including data extraction, preprocessing, model training, evaluation, prediction, and explanation—entirely with SQL, simplifying workflow construction across multiple databases and machine‑learning engines while supporting scalable execution on Kubernetes.

AI workflowKubernetesSQLFlow

0 likes · 19 min read

SQLFlow: End‑to‑End AI Workflow Construction Using SQL