[Model Request] Mosaic Pretrained Transformer (MPT)

[ThangPM]

⚙️ Request New Models

• Link to an existing implementation (e.g. Hugging Face/Github): https://huggingface.co/mosaicml/mpt-7b-instruct
• Is this model architecture supported by MLC-LLM? (the list of supported models) No.

Additional context

Hello, it's time to embrace MPT (Mosaic Pretrained Transformer) architecture in your open-source framework!

MPT offers:

1. Efficient scaling for large language models.
2. Improved performance across various NLP tasks.
3. Enhanced stability during training.
4. Open-source foundation for customization and innovation.

By incorporating MPT, you'll leverage a powerful, flexible architecture designed for the future of NLP.