solar-pro

Solar Pro Preview 採用增強版的深度向上擴展方法開發，將具有 140 億個參數的 Phi-3-medium 模型擴展到 220 億個，旨在於具有 80GB VRAM 的 GPU 上運行。訓練策略和資料集經過精心策劃，從而在 Phi-3-medium 之上實現了顯著的效能提升，尤其是在 MMLU-Pro 和 IFEval 等基準測試中，這些基準測試評估了模型的知識和指令遵循能力。

作為官方 Solar Pro 的預發布版本，Solar Pro Preview 在語言覆蓋範圍和最大上下文長度（4K）方面存在限制。儘管存在這些限制，但該模型仍以其效率和功能而著稱，並具有未來擴展以支援更多語言和功能的潛力。

官方版本的 Solar Pro 計劃於 2024 年 11 月發布，將包含擴展的語言支援和更長的上下文窗口。

參考文獻

部落格文章

Hugging Face

Solar Pro Preview is an advanced large language model (LLM) featuring 22 billion parameters, optimized to operate on a single GPU. It demonstrates superior performance compared to LLMs with fewer than 30 billion parameters and delivers results comparable to much larger models, such as Llama 3.1 with 70 billion parameters.

Developed using an enhanced version of the depth up-scaling method, Solar Pro Preview scales a Phi-3-medium model with 14 billion parameters to 22 billion, designed to run on a GPU with 80GB of VRAM. The training strategy and dataset have been meticulously curated, leading to significant performance improvements over Phi-3-medium, especially on benchmarks like MMLU-Pro and IFEval, which assess a model’s knowledge and instruction-following capabilities.

As a pre-release version of the official Solar Pro, Solar Pro Preview comes with limitations on language coverage and a maximum context length of 4K. Despite these restrictions, the model stands out for its efficiency and capability, with potential for future extensions to support more languages and functionalities.

The official version of Solar Pro, scheduled for release in November 2024, will include expanded language support and longer context windows.

## References

[Blog post](https://www.upstage.ai/products/solar-pro-preview)

[Hugging Face](https://huggingface.co/upstage/solar-pro-preview-instruct)

貼上、拖放或點擊以上傳圖片 (.png, .jpeg, .jpg, .svg, .gif)