File size: 607 Bytes
7a9ace6 cccf815 7a9ace6 cccf815 7a9ace6 cccf815 39379db ff9ea24 39379db 3957ef8 39379db |
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 35 36 37 |
---
language:
- en
- zh
- de
- fr
- es
- pt
- ru
- it
- ja
- ko
- vi
- ar
tags:
- pytorch
- text-generation
- causal-lm
- rwkv
license: apache-2.0
datasets:
- EleutherAI/pile
- togethercomputer/RedPajama-Data-1T
---
# RWKV-4 World
## Model Description
RWKV-4 trained on 100+ world languages.
How to use:
* use latest rwkv pip package (0.7.4+)
* use latest ChatRWKV v2/chat.py
* set pipeline = PIPELINE(model, "rwkv_vocab_v20230424") instead of 20B_tokenizer.json (EXACTLY AS WRITTEN HERE. "rwkv_vocab_v20230424" is included in rwkv 0.7.4+)
* Use Q/A or User/AI or Human/Bot prompt, instead of Bob/Alice
|