# English Assistant

NanoLM-1B-Instruct-v2 is a 1B-parameter small model fine-tuned on over 4 million high-quality instruction datasets, dedicated to exploring the potential of small models.

A quantized version of Meta-Llama-3.1-8B-Instruct, optimized to reduce disk space and GPU memory requirements, suitable for chat assistant scenarios in English business and research.

Meta-Llama-3-70B-Instruct-FP8 is a quantized version of Meta-Llama-3-70B-Instruct. It reduces disk size and GPU memory requirements through FP8 quantization while maintaining high performance. It is suitable for English business and research purposes.

Zephyr 141B-A39B is a large language model fine-tuned from Mixtral-8x22B-v0.1, trained using the ORPO alignment algorithm, designed to be a helpful assistant.