4 ms·
State-of-the-Art Ember embedding model for retrieval augmented generation
- enrike 3y agoModel has been trained on an extensive corpus of text pairs that encompass a broad spectrum of domains, including finance, science, medicine, law, and various others. During the training process, we incorporated techniques derived from the RetroMAE and SetFit research papers.