---
title: apar-13b
canonical_url: "https://www.modelscope.cn/models/ZhipuAI/apar-13b"
md_url: "https://www.modelscope.cn/models/ZhipuAI/apar-13b.md"
repository: ZhipuAI/apar-13b
last_updated: 2026-01-20
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - llama
architectures:
  - LlamaForCausalLM
library_name:
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy 0.9.1"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 81
stars: 0
---

# apar-13b

> apar-13b - ZhipuAI 在 ModelScope 开源的模型。[📃Paper: APAR: LLMs Can Do Auto-Parallel Auto-Regressive Decoding]

ZhipuAI/apar-13b 是 ModelScope 魔搭社区上的text-generation模型，可用 deploy_task text/emb、lmdeploy 0.9.1、lmdeploy_turbomind 0.9.1 部署。

- **Repository**: ZhipuAI/apar-13b
- **Tasks**: text-generation
- **Inference backends**: deploy_task text/emb, lmdeploy 0.9.1, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Downloads**: 81
- **Stars**: 0
- **Last updated**: 2026-01-20

Source: https://www.modelscope.cn/models/ZhipuAI/apar-13b

---

# APAR-13B

<center>
<p>
<a href="https://arxiv.org/abs/2401.06761" target="_blank">[📃Paper: APAR: LLMs Can Do Auto-Parallel Auto-Regressive Decoding]</a>
</p>
</center>

> The massive adoption of large language models (LLMs) demands efficient deployment strategies. However, the auto-regressive decoding process, which is fundamental to how most LLMs generate text, poses challenges to achieve efficient serving. In this work, we introduce a parallel auto-regressive generation method. By instruct-tuning on general domain data that contains hierarchical structures, we enable LLMs to independently plan their generation process and perform auto-parallel auto-regressive (APAR) generation, significantly reducing the number of generation steps. APAR alone can achieve up to 2x speed-up, and when combined with speculative decoding, the speed-up can reach up to 4x. In addition, APAR reduces the key-value cache consumption and attention computation during generation. This leads to a throughput increase of 20-70% and a latency reduce of 20-35% in high-throughput scenarios, compared to state-of-the-art serving frameworks.

**See our [paper](https://arxiv.org/abs/2401.06761) and [Github repo](https://github.com/THUDM/APAR) for details about the APAR-13B model.**