Update README.md
Browse files
README.md
CHANGED
|
@@ -1,67 +1,70 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-4.0
|
| 3 |
-
language:
|
| 4 |
-
- en
|
| 5 |
-
base_model:
|
| 6 |
-
- TeeZee/Orca-2-13b_flat
|
| 7 |
-
- NeverSleep/X-NoroChronos-13B
|
| 8 |
-
- NeverSleep/Noromaid-13b-v0.3
|
| 9 |
-
- KatyTheCutie/EstopianMaid-13B
|
| 10 |
-
- Undi95/MLewdBoros-L2-13B
|
| 11 |
-
- KoboldAI/LLaMA2-13B-Psyfighter2
|
| 12 |
-
- KoboldAI/LLaMA2-13B-Erebus-v3
|
| 13 |
-
library_name: transformers
|
| 14 |
-
tags:
|
| 15 |
-
- storywriting
|
| 16 |
-
- text adventure
|
| 17 |
-
- creative
|
| 18 |
-
- story
|
| 19 |
-
- writing
|
| 20 |
-
- fiction
|
| 21 |
-
- roleplaying
|
| 22 |
-
- rp
|
| 23 |
-
- mergekit
|
| 24 |
-
- merge
|
| 25 |
-
|
| 26 |
-
---
|
| 27 |
-
|
| 28 |
-

|
| 29 |
-
|
| 30 |
-
# Bionic-Vaquita-13B
|
| 31 |
-
|
| 32 |
-
In the same vein as the legendary [Psyonic-Cetacean-20B](https://huggingface.co/jebcarter/psyonic-cetacean-20B), I have attempted to create a 13B model that is equal parts creative and chaotic, while still remaining coherent enough for roleplaying purposes.
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
</br>
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
- sources:
|
| 60 |
-
- model: FallenMerick/
|
| 61 |
-
layer_range: [
|
| 62 |
-
- sources:
|
| 63 |
-
- model: FallenMerick/
|
| 64 |
-
layer_range: [
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
base_model:
|
| 6 |
+
- TeeZee/Orca-2-13b_flat
|
| 7 |
+
- NeverSleep/X-NoroChronos-13B
|
| 8 |
+
- NeverSleep/Noromaid-13b-v0.3
|
| 9 |
+
- KatyTheCutie/EstopianMaid-13B
|
| 10 |
+
- Undi95/MLewdBoros-L2-13B
|
| 11 |
+
- KoboldAI/LLaMA2-13B-Psyfighter2
|
| 12 |
+
- KoboldAI/LLaMA2-13B-Erebus-v3
|
| 13 |
+
library_name: transformers
|
| 14 |
+
tags:
|
| 15 |
+
- storywriting
|
| 16 |
+
- text adventure
|
| 17 |
+
- creative
|
| 18 |
+
- story
|
| 19 |
+
- writing
|
| 20 |
+
- fiction
|
| 21 |
+
- roleplaying
|
| 22 |
+
- rp
|
| 23 |
+
- mergekit
|
| 24 |
+
- merge
|
| 25 |
+
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+

|
| 29 |
+
|
| 30 |
+
# Bionic-Vaquita-13B
|
| 31 |
+
|
| 32 |
+
In the same vein as the legendary [Psyonic-Cetacean-20B](https://huggingface.co/jebcarter/psyonic-cetacean-20B), I have attempted to create a 13B model that is equal parts creative and chaotic, while still remaining coherent enough for roleplaying purposes.
|
| 33 |
+
</br>
|
| 34 |
+
Seven different Llama-2 13B models were hand-picked and merged via TIES to create three separate components for the final stack (to be made public upon request).
|
| 35 |
+
</br>
|
| 36 |
+
Emotional intelligence and coherency were the primary focus of the late-stage manual testing that led to selecting this model.
|
| 37 |
+
</br>
|
| 38 |
+
</br>
|
| 39 |
+
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
| 40 |
+
|
| 41 |
+
## Merge Details
|
| 42 |
+
### Merge Method
|
| 43 |
+
|
| 44 |
+
This model was merged using the passthrough merge method.
|
| 45 |
+
|
| 46 |
+
### Models Merged
|
| 47 |
+
|
| 48 |
+
The following models were included in the merge:
|
| 49 |
+
* FallenMerick/XNoroChronos-Orca2-Noromaid
|
| 50 |
+
* FallenMerick/Psyfighter2-Orca2-Erebus3
|
| 51 |
+
* FallenMerick/EstopianMaid-Orca2-MlewdBoros
|
| 52 |
+
|
| 53 |
+
### Configuration
|
| 54 |
+
|
| 55 |
+
The following YAML configuration was used to produce this model:
|
| 56 |
+
|
| 57 |
+
```yaml
|
| 58 |
+
slices:
|
| 59 |
+
- sources:
|
| 60 |
+
- model: FallenMerick/XNoroChronos-Orca2-Noromaid
|
| 61 |
+
layer_range: [0, 16]
|
| 62 |
+
- sources:
|
| 63 |
+
- model: FallenMerick/EstopianMaid-Orca2-MlewdBoros
|
| 64 |
+
layer_range: [16, 24]
|
| 65 |
+
- sources:
|
| 66 |
+
- model: FallenMerick/Psyfighter2-Orca2-Erebus3
|
| 67 |
+
layer_range: [24, 40]
|
| 68 |
+
merge_method: passthrough
|
| 69 |
+
dtype: bfloat16
|
| 70 |
+
```
|