Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- sophosympatheia/Midnight-Miqu-70B-v1.5
|
| 4 |
+
- NeverSleep/MiquMaid-v3-70B
|
| 5 |
+
- maywell/miqu-evil-dpo
|
| 6 |
+
- 152334H/miqu-1-70b-sf
|
| 7 |
+
library_name: transformers
|
| 8 |
+
tags:
|
| 9 |
+
- mergekit
|
| 10 |
+
- merge
|
| 11 |
+
license: other
|
| 12 |
+
---
|
| 13 |
+
# MiquSuperdark-70B-v1
|
| 14 |
+
|
| 15 |
+
**MiquSuperdark-70B-v1** is a merge of three of the most popular Miqu-derived models, along with Miqu itself. The goal of the merge is to create an strong, well-rounded chat model that picks up desirable traits from its constituent models without sacrificing intelligence.
|
| 16 |
+
|
| 17 |
+
This is a DARE Linear merge with the following composition:
|
| 18 |
+
- [sophosympatheia/Midnight-Miqu-70B-v1.5](https://huggingface.co/sophosympatheia/Midnight-Miqu-70B-v1.5) at weight 0.4
|
| 19 |
+
- [NeverSleep/MiquMaid-v3-70B](https://huggingface.co/NeverSleep/MiquMaid-v3-70B) at weight 0.2
|
| 20 |
+
- [maywell/miqu-evil-dpo](https://huggingface.co/maywell/miqu-evil-dpo) at weight 0.2
|
| 21 |
+
- [152334H/miqu-1-70b-sf](https://huggingface.co/152334H/miqu-1-70b-sf) at weight 0.2 (used as base model)
|
| 22 |
+
|
| 23 |
+
DARE Linear was chosen as the merge method based on [this HF discussion](https://huggingface.co/jukofyork/Dark-Miqu-70B/discussions/2), in which the creator of Midnight-Miqu says "*in my own testing I consistently got the best results from using a dare_linear merge when working with miqu models*".
|
| 24 |
+
|
| 25 |
+
## Merge Configuration
|
| 26 |
+
|
| 27 |
+
The following YAML configuration was used to produce this model:
|
| 28 |
+
|
| 29 |
+
```yaml
|
| 30 |
+
merge_method: dare_linear
|
| 31 |
+
base_model: /home/dylan/Documents/AI/merge/miqu-1-70b-sf
|
| 32 |
+
models:
|
| 33 |
+
- model: /media/dylan/SanDisk/LLMs/Midnight-Miqu-70B-v1.5
|
| 34 |
+
parameters:
|
| 35 |
+
weight: 0.4
|
| 36 |
+
- model: /home/dylan/Documents/AI/merge/miqu-1-70b-sf
|
| 37 |
+
parameters:
|
| 38 |
+
weight: 0.2
|
| 39 |
+
- model: /media/dylan/SanDisk/LLMs/miqu-evil-dpo/
|
| 40 |
+
parameters:
|
| 41 |
+
weight: 0.2
|
| 42 |
+
- model: /home/dylan/Documents/AI/merge/MiquMaid-v3-70B
|
| 43 |
+
parameters:
|
| 44 |
+
weight: 0.2
|
| 45 |
+
dtype: float16
|
| 46 |
+
tokenizer_source: model:/home/dylan/Documents/AI/merge/miqu-1-70b-sf
|
| 47 |
+
```
|
| 48 |
+
|
| 49 |
+
The tokenizer is copied from the base model [152334H/miqu-1-70b-sf](https://huggingface.co/152334H/miqu-1-70b-sf).
|