---
title: rwkv-7B-world-novel-128k
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/rwkv-7B-world-novel-128k"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/rwkv-7B-world-novel-128k.md"
repository: AI-ModelScope/rwkv-7B-world-novel-128k
last_updated: 2023-10-28
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
library_name:
  - pytorch
frameworks:
  - Pytorch
downloads: 2535
stars: 6
---

# rwkv-7B-world-novel-128k

> rwkv-7B-world-novel-128k - AI-ModelScope 在 ModelScope 开源的模型。RWKV 7B World 128k for novel writing

AI-ModelScope/rwkv-7B-world-novel-128k 是 ModelScope 魔搭社区上的text-generation模型，采用 apache-2.0 许可。

- **Repository**: AI-ModelScope/rwkv-7B-world-novel-128k
- **License**: apache-2.0
- **Tasks**: text-generation
- **Downloads**: 2535
- **Stars**: 6
- **Last updated**: 2023-10-28

Source: https://www.modelscope.cn/models/AI-ModelScope/rwkv-7B-world-novel-128k

---

# RWKV 7B World 128k for novel writing

We proudly announce this is the world first **128k context** model based on RWKV architecture today, 2023-08-10.

With RWKV world tokenizer,multi-langs have 1:1 tokenization ratio ,one word to one token.
(https://github.com/BlinkDL/ChatRWKV/blob/2a13ddecd81f8fd615b6da3a8f1091a594689e30/tokenizer/rwkv_tokenizer.py#L163)


# How to train infinte context model?

This model trained with instructions datasets and chinese web novel and tradition wuxia, 
more trainning details would be updated.

Tested to summary 85k tokens  to 5 keypoints ,can find conversation files in example folders ,more cases are coming.

Full finetuned using this repo to train 128k context model , 4*A800 40hours with 1.3B tokens.
https://github.com/SynthiaDL/TrainChatGalRWKV/blob/main/train_world.sh

![QQ图片20230810153529.jpg](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/d8ekmc4Lfhy2lYEdrRKXz.jpeg)

# How to Test?

Using RWKV Runner https://github.com/josStorer/RWKV-Runner  to test this model, only need 16G vram to run fp16 or 8G vram fp16i8, use temp 0.1-0.2 topp 0.7 for more precise answer ,temp between 1-2.x is more creatively.
![微信截图_20230810162303.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/Ww45-WMngl4Jyt1OZDAa_.png)


![image.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/5zDQVbGb-fX8Y8h98tUF0.png)

![微信截图_20230810142220.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/u2wA-l1UcW-Mt9KIoa_4q.png)

![4UYBX4RA0%8PA{1YSSK)AVW.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/gzr8Yt4JRkBz31-ieRSOE.png)

![QQ图片20230810143840.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/LgEjfHJ7XD7PlGM9b3RAf.png)

![image.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/b_6KCBdZKW7Q7HwipxE-l.png)

85k tokens test
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/F9unOJfhmJPXsciPHLsrl.png)

![image.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/35j5C1QD_4cO-AjfxV7tl.png)

![微信截图_20230810201844.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/5dVQrJxg05C0ww7_AVhrW.png)

![83be699bab815e4396254eb5e855090.png](https://cdn-uploads.huggingface.co/production/uploads/6176b32847ee6431f632981e/yde7ZcsxecbrQUIFvDtUs.png)