File size: 1,011 Bytes
f65da4f a3c2667 8c6855e f65da4f 31fab84 2ebd3a9 31fab84 2ebd3a9 31fab84 2ebd3a9 31fab84 2ebd3a9 31fab84 86088f5 31fab84 2ebd3a9 31fab84 ca3882e 31fab84 ca3882e 31fab84 ca3882e 31fab84 ca3882e |
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 |
---
datasets:
- PygmalionAI/PIPPA
- ludis/geepeetee4
- lemonilia/LimaRP
---
## Prompting
https://rentry.org/tsukasa13b - reccomended prompts and gen settings
The current model version has been trained on prompts using three different roles, which are denoted by the following tokens: `<|system|>`, `<|user|>` and `<|model|>`.
The `<|system|>` prompt can be used to inject out-of-channel information behind the scenes, while the `<|user|>` prompt should be used to indicate user input. The `<|model|>` token should then be used to indicate that the model should generate a response. These tokens can happen multiple times and be chained up to form a conversation history.
## Training
base model (llama-2-13b-hf)
tuned on koishi dataset (commit c83d922) for 1 epoch
then tuned on pippa dataset (commit 6412b0c) for 1 epoch
then tuned on geepeetee4 dataset (commit c83d922) for 1 epoch
then tuned on limarp (without ponyville, lolicit, and all the fallen subsets. Version 2023-09-14) for 2 epochs
|