---
title: wizard-rwkv-world-7B-ctx32k
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/wizard-rwkv-world-7B-ctx32k"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/wizard-rwkv-world-7B-ctx32k.md"
repository: AI-ModelScope/wizard-rwkv-world-7B-ctx32k
last_updated: 2023-10-28
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
library_name:
  - pytorch
frameworks:
  - Pytorch
downloads: 28
stars: 0
---

# wizard-rwkv-world-7B-ctx32k

> wizard-rwkv-world-7B-ctx32k - AI-ModelScope 在 ModelScope 开源的模型。This model is finetuned based on RWKV4 world 7B english model, using wizard datset to fit in 32k ctx.

AI-ModelScope/wizard-rwkv-world-7B-ctx32k 是 ModelScope 魔搭社区上的text-generation模型，采用 apache-2.0 许可。

- **Repository**: AI-ModelScope/wizard-rwkv-world-7B-ctx32k
- **License**: apache-2.0
- **Tasks**: text-generation
- **Downloads**: 28
- **Stars**: 0
- **Last updated**: 2023-10-28

Source: https://www.modelscope.cn/models/AI-ModelScope/wizard-rwkv-world-7B-ctx32k

---

This model is finetuned based on RWKV4 world 7B english model, using wizard datset to fit in 32k ctx.

should do some complex instruction and can feed in 32k token, but as we know wizard dataset only have max 4K length per sample ,

i'll do more test on long prompt and generation

use https://github.com/josStorer/RWKV-Runner to run this model.



![QQ截图20230729111155.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/OCapn4mFYvXAXU1PfIwf8.png)

![image.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/mScVGl8mhETw8qq45JeoM.png)