---
title: Qwen1.5-MoE-A2.7B
canonical_url: "https://www.modelscope.cn/models/Qwen/Qwen1.5-MoE-A2.7B"
md_url: "https://www.modelscope.cn/models/Qwen/Qwen1.5-MoE-A2.7B.md"
repository: Qwen/Qwen1.5-MoE-A2.7B
chinese_name: "千问1.5-MoE-A2.7B"
last_updated: 2024-04-18
license: other
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen2_moe
architectures:
  - Qwen2MoeForCausalLM
parameters: 14.3B
tensor_type:
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy 0.9.1"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 16666
stars: 23
tags:
  - pretrained
  - moe
---

# Qwen1.5-MoE-A2.7B

> Qwen1.5-MoE-A2.7B - Qwen 在 ModelScope 开源的模型。Qwen1.5-MoE is a transformer-based MoE decoder-only language model pretrained on a large amount of data.

Qwen/Qwen1.5-MoE-A2.7B 是 ModelScope 魔搭社区上的 14.3B 参数text-generation模型，采用 other 许可，可用 deploy_task text/emb、lmdeploy 0.9.1、lmdeploy_turbomind 0.9.1 部署。

- **Repository**: Qwen/Qwen1.5-MoE-A2.7B
- **License**: other
- **Tasks**: text-generation
- **Parameters**: 14.3B
- **Inference backends**: deploy_task text/emb, lmdeploy 0.9.1, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Tags**: pretrained, moe
- **Downloads**: 16666
- **Stars**: 23
- **Last updated**: 2024-04-18

Source: https://www.modelscope.cn/models/Qwen/Qwen1.5-MoE-A2.7B

---

# Qwen1.5-MoE-A2.7B


## Introduction

Qwen1.5-MoE is a transformer-based MoE decoder-only language model pretrained on a large amount of data. 

For more details, please refer to our [blog post](https://qwenlm.github.io/blog/qwen-moe/) and [GitHub repo](https://github.com/QwenLM/Qwen1.5).

## Model Details
Qwen1.5-MoE employs Mixture of Experts (MoE) architecture, where the models are upcycled from dense language models. For instance, `Qwen1.5-MoE-A2.7B` is upcycled from `Qwen-1.8B`. It has 14.3B parameters in total and 2.7B activated parameters during runtime, while achieving comparable performance to `Qwen1.5-7B`, it only requires 25% of the training resources. We also observed that the inference speed is 1.74 times that of `Qwen1.5-7B`.

## Requirements
The code of Qwen1.5-MoE has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error:
```
KeyError: 'qwen2_moe'.
```

## Usage

We do not advise you to use base language models for text generation. Instead, you can apply post-training, e.g., SFT, RLHF, continued pretraining, etc., on this model.