File size: 1,598 Bytes
fdaf88d 460d03e fdaf88d |
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 |
---
license: cc-by-nc-4.0
language:
- en
---
ExllamaV2 version of the model created by [Sao10K](https://huggingface.co./Sao10K)!
Original Model https://huggingface.co./Sao10K/WinterGoddess-1.4x-70B-L2
Requires ExllamaV2, which is being developed by turboderp https://github.com/turboderp/exllamav2 under an MIT license.
-----------------
Winter Goddess - A 70B L2 Model for General use, or for Roleplay.
I wanted a Smart Model that is Capable of following Instructions, while being able to (e)RP effectively. Sort of like 1.3, but better.
I merged some models as a base, and had tuned on top of it afterwards.
I personally think this mogs Euryale 1.3, but ymmv.
***
For Transparency's Sake:
Models Used:
<br> Platypus2-70B-instruct
<br> Lila-70B
<br> SunsetBoulevard (at roughly 0.1 weight, boosting coherency)
<br> Private De-alignment LoRA on top.
why does it show mergekit in the safetensors.index metadata? -> I used DARE method to merge the 3 models. Then Axolotl qLoRA. then used lora-merge, copying the files of the base merged model because they didn't save to the new one, only the .safetensor files got saved.
***
Prompt Format - Alpaca
```
### Instruction:
<Prompt>
### Response:
```
OR
```
### Instruction:
<Prompt>
### Input:
<Insert Context Here>
### Response:
```
***
<br> 42. A 25-year-old female has been struck in the right eye with a pipe. She has a ruptured right globe, an orbital fracture and no other obvious injury. You should bandage:
<br> A) The right eye tightly
<br> B) Both eyes loosely
<br> C) The right eye loosely
<br> D) Both eyes tightly
|