Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,46 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
---
|
5 |
+
license: apache-2.0
|
6 |
+
---
|
7 |
+
# Model url: https://huggingface.co/TimeMobius/Mobius-RWKV-r5-chat-12B-8k
|
8 |
+
Considering the long context required for training from scratch, we decided to retrain the r5 12B model from 8k.
|
9 |
+
This model exhibits lower diversity compared to its predecessor, but it excels in following instructions and logical understanding. It is possible to utilize both models simultaneously as multi-agents, each performing a different task.
|
10 |
+
|
11 |
+
# Mobius RWKV r5 chat 12B 8k
|
12 |
+
Mobius is a RWKV v5.2 arch chat model, benifit from [Matrix-Valued States and Dynamic Recurrence](https://arxiv.org/abs/2404.05892)
|
13 |
+
|
14 |
+
## Introduction
|
15 |
+
|
16 |
+
Mobius is a RWKV v5.2 arch model, a state based RNN+CNN+Transformer Mixed language model pretrained on a certain amount of data.
|
17 |
+
In comparison with the previous released Mobius, the improvements include:
|
18 |
+
|
19 |
+
* Only 24G Vram to run this model locally with fp16;
|
20 |
+
* Significant performance improvement;
|
21 |
+
* Multilingual support ;
|
22 |
+
* Stable support of 128K context length.
|
23 |
+
* Base model [Mobius-mega-12B-128k-base](https://huggingface.co/TimeMobius/Moibus-mega-12B-128k-base)
|
24 |
+
|
25 |
+
|
26 |
+
## Usage
|
27 |
+
We encourage you use few shots to use this model, Desipte Directly use User: xxxx\n\nAssistant: xxx\n\n is really good too, Can boost all potential ability.
|
28 |
+
|
29 |
+
Recommend Temp and topp: 0.7 0.6/1 0.3/1.5 0.3/0.2 0.8
|
30 |
+
|
31 |
+
## More details
|
32 |
+
Mobius 12B 128k based on RWKV v5.2 arch, which is leading state based RNN+CNN+Transformer Mixed large language model which focus opensouce community
|
33 |
+
* 10~100 trainning/inference cost reduce;
|
34 |
+
* state based,selected memory, which mean good at grok;
|
35 |
+
* community support.
|
36 |
+
|
37 |
+
## requirements
|
38 |
+
24G vram to run fp16, 12G for int8, 6G for nf4 with Ai00 server.
|
39 |
+
|
40 |
+
* [RWKV Runner](https://github.com/josStorer/RWKV-Runner)
|
41 |
+
* [Ai00 server](https://github.com/cgisky1980/ai00_rwkv_server)
|
42 |
+
|
43 |
+
## future plan
|
44 |
+
If you need a HF version let us know
|
45 |
+
|
46 |
+
[Mobius-Chat-12B-128k](https://huggingface.co/TimeMobius/Mobius-Chat-12B-128k)
|