{"name":"jmagly-training","owner":{"name":"ClaudePluginHub"},"plugins":[{"name":"jmagly-training","source":{"source":"github","repo":"jmagly/aiwg-training"},"description":"Corpus-to-dataset pipeline for AI training data curation. Ingests sources, synthesizes examples, generates preference pairs, applies decontamination, and exports to Alpaca/ShareGPT/ChatML/JSONL/Parquet with provenance and reproducibility. Grounded in 485 research REFs covering DPO/KTO/ORPO/SimPO, Self-Instruct/Evol/Orca/Phi/PersonaHub/STaR/ReST, Model Collapse guard, Datasheets/Model Cards/Data Statements, HF Datasets/Arrow+Parquet.","version":"1.0.0","strict":true,"keywords":["training-data","fine-tuning","dpo","kto","orpo","alpaca","sharegpt","dataset-versioning","decontamination","synthetic-data","datasheet","model-card","aiwg","fine-tuning"],"category":"deployment"}]}