Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: llama3
|
| 3 |
+
language:
|
| 4 |
+
- ja
|
| 5 |
+
- en
|
| 6 |
+
tags:
|
| 7 |
+
- moe
|
| 8 |
+
- japanese
|
| 9 |
+
- sql
|
| 10 |
+
---
|
| 11 |
+
### モデルの説明(English explanation is below.)
|
| 12 |
+
このモデルは、MergeKitツールを使用して作成されたMixture of Experts (MoE) 言語モデルをGGUF形式で量子化したものです。
|
| 13 |
+
量子化していないものは [こちら](https://huggingface.co/keitokei1994/swallow-3-8B-sqlcoder-2x8B) 。
|
| 14 |
+
|
| 15 |
+
### モデルの詳細
|
| 16 |
+
- **モデル名**: swallow-3-8B-sqlcoder-2x8B-GGUF
|
| 17 |
+
- **モデルアーキテクチャ**: Mixture of Experts (MoE)
|
| 18 |
+
- **ベースモデル**:
|
| 19 |
+
- [tokyotech-llm/Llama-3-Swallow-8B-Instruct-v0.1](https://huggingface.co/tokyotech-llm/Llama-3-Swallow-8B-Instruct-v0.1)
|
| 20 |
+
- [defog/llama-3-sqlcoder-8b](https://huggingface.co/defog/llama-3-sqlcoder-8b)
|
| 21 |
+
- **マージツール**: MergeKit
|
| 22 |
+
このMoEモデルは、Llama-3-Swallow-8B-Instruct-v0.1の日本語能力とLlama-3-sqlcoder-8bのSQL生成能力を組み合わせることで、より強力で多機能な言語モデルを目指しています。
|
| 23 |
+
#### 特徴
|
| 24 |
+
- 日本語と英語の両方に対応
|
| 25 |
+
- Llama-3-Swallow-8B-Instruct-v0.1による優れた日本語処理能力
|
| 26 |
+
- Llama-3-sqlcoder-8bによる高度なSQL生成と処理能力
|
| 27 |
+
#### 要求スペック
|
| 28 |
+
Q4_K_M量子化モデルであれば、RTX3060 12GBでフルロード可能です。
|
| 29 |
+
筆者はWSL2やGoogle Colaboratotry Proでの作成後、Llama.cppとLMstudioにて動作確認を行っています。
|
| 30 |
+
|
| 31 |
+
---
|
| 32 |
+
### Model Description
|
| 33 |
+
This model is a Mixture of Experts (MoE) language model created using the MergeKit tool.
|
| 34 |
+
The gguf version can be found [こちら](https://huggingface.co/keitokei1994/swallow-3-8B-sqlcoder-2x8B).
|
| 35 |
+
### Model Details
|
| 36 |
+
- **Model Name**: swallow-3-8B-sqlcoder-2x8B-GGUF
|
| 37 |
+
- **Model Architecture**: Mixture of Experts (MoE)
|
| 38 |
+
- **Base Models**:
|
| 39 |
+
- [tokyotech-llm/Llama-3-Swallow-8B-Instruct-v0.1](https://huggingface.co/tokyotech-llm/Llama-3-Swallow-8B-Instruct-v0.1)
|
| 40 |
+
- [defog/llama-3-sqlcoder-8b](https://huggingface.co/defog/llama-3-sqlcoder-8b)
|
| 41 |
+
- **Merge Tool**: MergeKit
|
| 42 |
+
This MoE model aims to create a more powerful and versatile language model by combining the Japanese language capabilities of Llama-3-Swallow-8B-Instruct-v0.1 with the SQL generation abilities of Llama-3-sqlcoder-8b.
|
| 43 |
+
#### Features
|
| 44 |
+
- Support for both Japanese and English languages
|
| 45 |
+
- Excellent Japanese processing capabilities from Llama-3-Swallow-8B-Instruct-v0.1
|
| 46 |
+
- Advanced SQL generation and processing capabilities from Llama-3-sqlcoder-8b
|
| 47 |
+
#### System Requirements
|
| 48 |
+
If using the Q4_K_M quantized model, it can be fully loaded on an RTX3060 12GB.
|
| 49 |
+
The author has created the model using WSL2 and Google Colaboratory Pro, and has tested it using Llama.cpp and LMstudio.
|