File size: 1,699 Bytes
97de836 84027e2 c9f5a43 97de836 7fd139a 97de836 7fd139a 97de836 7fd139a 97de836 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 |
---
license: mit
pipeline_tag: text-generation
tags:
- cortex.cpp
---
## Overview
**Microsoft** developed and released the [Phi-3.5](https://huggingface.co/microsoft/Phi-3.5-mini-instruct) model, a state-of-the-art large language model built upon the Phi-3 architecture. With its focus on high-quality, reasoning-dense data, this model represents a significant advancement in instruction-tuned language models. Phi-3.5 has been fine-tuned through supervised learning, proximal policy optimization (PPO), and direct preference optimization (DPO) to ensure precise instruction following and robust safety measures. Supporting a 128K token context length, the model demonstrates exceptional performance in tasks requiring extended context understanding and complex reasoning. The model's training data consists of synthetic datasets and carefully filtered publicly available web content, inheriting the high-quality foundation established in the Phi-3 series.
## Variants
| No | Variant | Cortex CLI command |
| --- | --- | --- |
| 1 | [Phi-3.5-3b](https://huggingface.co/cortexso/phi-3.5/tree/3b) | `cortex run phi-3.5:3b` |
## Use it with Jan (UI)
1. Install **Jan** using [Quickstart](https://jan.ai/docs/quickstart)
2. Use in Jan model Hub:
```bash
cortexso/phi-3.5
```
## Use it with Cortex (CLI)
1. Install **Cortex** using [Quickstart](https://cortex.jan.ai/docs/quickstart)
2. Run the model with command:
```bash
cortex run phi-3.5
```
## Credits
- **Author:** Microsoft
- **Converter:** [Homebrew](https://www.homebrew.ltd/)
- **Original License:** [License](https://choosealicense.com/licenses/mit/)
- **Papers:** [Phi-3.5 Paper](https://arxiv.org/abs/2404.14219) |