File size: 2,447 Bytes
d48d1c5 |
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 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
---
language:
- en
- zh
- fr
- es
- de
- pt
- ru
- it
- ja
- ko
- vi
- ar
tags:
- pytorch
- text-generation
- causal-lm
- rwkv
license: apache-2.0
datasets:
- cerebras/SlimPajama-627B
- EleutherAI/pile
- bigcode/starcoderdata
- oscar-corpus/OSCAR-2301
---
# RWKV-5 World
Use rwkv pip package 0.8.22+ for RWKV-5 inference: https://pypi.org/project/rwkv/ (pipeline = PIPELINE(model, "rwkv_vocab_v20230424") for rwkv-world models)
Online 7B Demo: https://huggingface.co/spaces/BlinkDL/RWKV-Gradio-2
Online 1.5B Demo: https://huggingface.co/spaces/BlinkDL/RWKV-Gradio-1
GUI: https://github.com/josStorer/RWKV-Runner (see Releases)
Convert to HF formet: https://github.com/BBuf/RWKV-World-HF-Tokenizer
For developer: https://github.com/BlinkDL/ChatRWKV/blob/main/API_DEMO_CHAT.py
https://github.com/BlinkDL/ChatRWKV/blob/main/RWKV_v5_demo.py
How it works: https://twitter.com/BlinkDL_AI/status/1685230712247795713
https://www.rwkv.com/
## Model Description
RWKV-5 trained on 100+ world languages (70% English, 15% multilang, 15% code).
World = Some_Pile + Some_SlimPajama + Some_StarCoder + Some_OSCAR + All_Wikipedia + All_ChatGPT_Data_I_can_find
RWKV-5 training: set --my_testing "r2r4" in latest RWKV-LM v4neo: https://github.com/BlinkDL/RWKV-LM
World v1 = 0.59T tokens
World v2 = 1.12T tokens
Imagine what happens when we use more data :)
Recommended fine-tuning format (use \n for newlines):
```
User: xxxxxxxxxxxxxxx
Assistant: xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
User: xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
Assistant: xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
```
A good chat prompt (better replace \n\n in xxx to \n, such that there will be no newlines in xxx):
```
User: hi
Assistant: Hi. I am your assistant and I will provide expert full response in full details. Please feel free to ask any question and I will always answer it.
User: xxx
Assistant:
```
QA prompt (better replace \n\n in xxx to \n, such that there will be no newlines in xxx):
```
Question: xxx
Answer:
```
and
```
Instruction: xxx
Input: xxx
Response:
```
!!! There should not be any space after your final ":" or you will upset the tokenizer and see non-English reponse !!!
!!! There should not be any space after your final ":" or you will upset the tokenizer and see non-English reponse !!!
!!! There should not be any space after your final ":" or you will upset the tokenizer and see non-English reponse !!!
|