Update README.md

README.md

@@ -1,40 +1,78 @@
----
-base_model: []
-library_name: transformers
-tags:
-- mergekit
-- merge
-
----
-# Antler-Straling-dare-ties-2
-
-This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
-
-## Merge Details
-### Merge Method
-
-This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using ./Antler-7B-RP-v3-Model-Stock as a base.
-
-### Models Merged
-
-The following models were included in the merge:
-* ./Japanese-Starling-ChatV-7B-RP-Model-Stock
-
-### Configuration
-
-The following YAML configuration was used to produce this model:
-
-```yaml
-models:
-  - model: ./Antler-7B-RP-v3-Model-Stock
-    # no parameters necessary for base model
-  - model: ./Japanese-Starling-ChatV-7B-RP-Model-Stock # follow user intent
-    parameters:
-      density: 1
-      weight: 0.7
-merge_method: dare_ties
-base_model: ./Antler-7B-RP-v3-Model-Stock
-dtype: bfloat16
-tokenizer_source: union
-
-```
+---
+base_model:
+- Aratako/Antler-7B-RP-v3
+- Aratako/Japanese-Starling-ChatV-7B-RP
+- senseable/WestLake-7B-v2
+- SanjiWatsuki/Kunoichi-DPO-v2-7B
+- SanjiWatsuki/Silicon-Maid-7B
+- SanjiWatsuki/Loyal-Macaroni-Maid-7B
+library_name: transformers
+tags:
+- mergekit
+- merge
+- not-for-all-audiences
+- nsfw
+language:
+- ja
+license: apache-2.0
+---
+# AntlerStar-RP
+[GGUF版はこちら/Click here for the GGUF version](https://huggingface.co/Aratako/AntlerStar-RP-GGUF)
+
+## 概要
+This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
+
+[Aratako/Antler-7B-RP-v3](https://huggingface.co/Aratako/Antler-7B-RP-v3)と[Aratako/Japanese-Starling-ChatV-7B-RP](https://huggingface.co/Aratako/Japanese-Starling-ChatV-7B-RP)の2つのモデルをベースにマージして作成したロールプレイ用モデルです。
+
+## マージの詳細
+まず、[Aratako/Antler-7B-RP-v3](https://huggingface.co/Aratako/Antler-7B-RP-v3)と[Aratako/Japanese-Starling-ChatV-7B-RP](https://huggingface.co/Aratako/Japanese-Starling-ChatV-7B-RP)の2モデルに対し、以下4モデルのChat Vectorを0.5倍して加算し、各4種類、計8種類のChat Vector加算モデルを作成しました。
+
+- [senseable/WestLake-7B-v2](https://huggingface.co/senseable/WestLake-7B-v2)
+- [SanjiWatsuki/Kunoichi-DPO-v2-7B](https://huggingface.co/SanjiWatsuki/Kunoichi-DPO-v2-7B)
+- [SanjiWatsuki/Silicon-Maid-7B](https://huggingface.co/SanjiWatsuki/Silicon-Maid-7B)
+- [SanjiWatsuki/Loyal-Macaroni-Maid-7B](https://huggingface.co/SanjiWatsuki/Loyal-Macaroni-Maid-7B)
+
+次に、このChat Vector加算によってできた各4モデルと元のモデルを、それぞれModel Stockという手法を用い以下のようなconfigを使ってmergekitでマージし、2つのモデルを作成しました。
+
+```yaml
+models:
+  - model: ./Antler-7B-RP-v3
+  - model: ./Antler-7B-RP-v3-WestLake-ChatVector
+  - model: ./Antler-7B-RP-v3-Kunoichi-ChatVector
+  - model: ./Antler-7B-RP-v3-SiliconMaid-ChatVector
+  - model: ./Antler-7B-RP-v3-LoyalMacaroniMaid-ChatVector
+merge_method: model_stock
+base_model: ./Antler-7B-RP-v3
+dtype: bfloat16
+tokenizer_source: union
+```
+
+```yaml
+models:
+  - model: ./Japanese-Starling-ChatV-7B-RP
+  - model: ./Japanese-Starling-ChatV-7B-RP-WestLake-ChatVector
+  - model: ./Japanese-Starling-ChatV-7B-RP-Kunoichi-ChatVector
+  - model: ./Japanese-Starling-ChatV-7B-RP-SiliconMaid-ChatVector
+  - model: ./Japanese-Starling-ChatV-7B-RP-LoyalMacaroniMaid-ChatVector
+merge_method: model_stock
+base_model: ./Japanese-Starling-ChatV-7B-RP
+dtype: bfloat16
+tokenizer_source: union
+```
+
+最後に、この2つのモデルを[DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708)という手法でmergekitを用いてマージしました。
+
+```yaml
+models:
+  - model: ./Antler-7B-RP-v3-Model-Stock
+    # no parameters necessary for base model
+  - model: ./Japanese-Starling-ChatV-7B-RP-Model-Stock # follow user intent
+    parameters:
+      density: 1
+      weight: 0.7
+merge_method: dare_ties
+base_model: ./Antler-7B-RP-v3-Model-Stock
+dtype: bfloat16
+tokenizer_source: union
+
+```