1st Multilingual Model Workshop – Training an instruction tuned and aligned LLM for Indian languages

This talk covered challenges in creating high-quality pre-training, finetuning, and preference optimization datasets. Pratyush shared experiences in training the models and discussed benchmarks and accuracy evaluation. Pratyush also announced the availability of the models at a hosted inference endpoint and discussed how the model brings a significant efficiency gain for Indian languages over existing models (including GPT-x).