- Added extra args for RAG: - `split_text_character` - `split_text_n`
0.61.00.61.0
Added
- Added support for RAG models (in `Seq2Seq`) - docs will be updated soon - Added support for Huggingface Datasets library for memory efficient training. Currently supports: - Classification (all) - NER - Language Modeling - Seq2Seq - T5 - QA (Note that HF Datasets might not always work with QAModel)