File size: 1,808 Bytes
327ebbf d8b1e8b 327ebbf 34eddd5 327ebbf |
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 |
---
language:
- en
library_name: transformers
pipeline_tag: text-generation
tags:
- llama
- llama-2
license: agpl-3.0
---
# Model Card: Nous-Hermes-limarp-l2-13B
This is a merge between:
- [Nous Hermes Llama 2 13b](https://huggingface.co/NousResearch/Nous-Hermes-Llama2-13b)
- [LimaRP llama 2 Lora](https://huggingface.co/lemonilia/limarp-llama2) from July 28, 2023 at a weight of 0.66.
Quantizations provided by myself:
- [GPTQ](https://huggingface.co/kingbri/Nous-Hermes-limarp-l2-13B-GPTQ)
The merge was performed using [ez-trainer](https://github.com/CoffeeVampir3/ez-trainer) by CoffeeVampire/Blackroot
The intended objective is to combine NH-2's reasoning and instruction-following capabilities with LimaRP's character roleplay capabilities.
The LimaRP LoRA was merged at a weight of 0.66 since a merge of 1.00 destroyed most nuances of a character's personality due to the LoRA being too strong on the base model.
There still may be edge cases. If so, please report them and the lora weight will be dropped.
## Usage:
Since this is a merge between NH-2 and LimaRP, the following instruction formats should work:
Alpaca 2:
```
### Instruction:
<prompt>
### Response:
<leave a newline blank for model to respond>
```
LimaRP instruction format:
```
<<SYSTEM>>
<character card and system prompt>
<<USER>>
<prompt>
<<AIBOT>>
<leave a newline blank for model to respond>
```
## Bias, Risks, and Limitations
The model will show biases similar to those observed in niche roleplaying forums on the Internet, besides those exhibited by the base model. It is not intended for supplying factual information or advice in any form.
## Training Details
This model is merged and can be reproduced using the tools mentioned above. Please refer to all provided links for extra model-specific details. |