Hunyuan 7B Instruct 0124

🚀 Hunyuan-7B Model

The 7B models released by Hunyuan use better data allocation and training, achieving a good balance between computing and performance, and standing out among large language models.

✨ Features

• The 7B models released by Hunyuan this time, Hunyuan-7B-Pretrain-0124 and Hunyuan-7B-Instruct-0124, have strong performance and are currently one of the strongest Chinese 7B Dense models.
• Extended long text capability to 256K and utilizes Grouped Query Attention (GQA).
• This open - source release offers two inference backend options tailored for the Hunyuan - 7B model: the popular vLLM - backend and the TensorRT - LLM Backend. Initially, the vLLM solution is open - sourced, and the TRT - LLM solution will be released soon.
• The Hunyuan - 7B open - source model is fully compatible with the Hugging Face format, enabling fine - tuning using the hf - deepspeed framework.

📚 Documentation

Model Introduction

The 7B models released by Hunyuan this time: Hunyuan-7B-Pretrain-0124 and Hunyuan-7B-Instruct-0124, use better data allocation and training, have strong performance, and have achieved a good balance between computing and performance. It stands out from many large - scale language models and is currently one of the strongest Chinese 7B Dense models.

Introduction to Technical Advantages

Model

• Extended long text capability to 256K and utilizes Grouped Query Attention (GQA).

Inference Framework

This open - source release offers two inference backend options tailored for the Hunyuan - 7B model: the popular vLLM - backend and the TensorRT - LLM Backend. In this release, we are initially open - sourcing the vLLM solution, with plans to release the TRT - LLM solution in the near future.

Training Framework

The Hunyuan - 7B open - source model is fully compatible with the Hugging Face format, enabling researchers and developers to perform model fine - tuning using the hf - deepspeed framework. Learn more: Tencent - Hunyuan - Large.

Related News

• 2025.1.24 We have open - sourced Hunyuan - 7B - Pretrain - 0124 and Hunyuan - 7B - Instruct - 0124 on Hugging Face.

Benchmark

Note: The following benchmarks are evaluated by TRT - LLM - backend

Hunyuan - 7B - Pretrain

Hunyuan - 7B - Instruct

🚀 Quick Start

You can refer to the content in Tencent - Hunyuan - Large to get started quickly. The training and inference code can use the version provided in this github repository.

Inference Framework

This open - source release offers two inference backend options tailored for the Hunyuan - 7B model: the popular vLLM - backend and the TensorRT - LLM Backend. In this release, we are initially open - sourcing the vLLM solution, with plans to release the TRT - LLM solution in the near future.

Inference Performance

This section presents the efficiency test results of deploying various models using vLLM, including inference speed (tokens/s) under different batch sizes.

📄 License

The model is under the tencent - license.

Contact Us

If you would like to leave a message for our R & D and product teams, welcome to contact our open - source team. You can also contact us via email (hunyuan_opensource@tencent.com).