Update README.md
Browse files
README.md
CHANGED
|
@@ -12,13 +12,12 @@ tags:
|
|
| 12 |
|
| 13 |
This model was created to use RAG. [bge-m3](https://huggingface.co/BAAI/bge-m3) based model. This embedding model for Dungeon & Fighter data.
|
| 14 |
|
| 15 |
-
Trained data, [Kakao_brain](https://github.com/kakaobrain/kor-nlu-datasets) koSTS, koNLI Data
|
| 16 |
|
| 17 |
Fianl Model, bge-m3-ko_DNF is Mixed bge-m3 and trained model. Mixed reason detail read bge-m3 [Page](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune)
|
| 18 |
|
| 19 |
While the overall similarity has decreased, the discriminatory power between queries (questions) and chunks (answer search space) has increased, allowing for more accurate parsing of Dungeon & Fighters information.
|
| 20 |
|
| 21 |
-
|
| 22 |
<!--- Describe your model here -->
|
| 23 |
|
| 24 |
## Usage (Sentence-Transformers)
|
|
|
|
| 12 |
|
| 13 |
This model was created to use RAG. [bge-m3](https://huggingface.co/BAAI/bge-m3) based model. This embedding model for Dungeon & Fighter data.
|
| 14 |
|
| 15 |
+
Trained data, [Kakao_brain](https://github.com/kakaobrain/kor-nlu-datasets) koSTS, koNLI Data and then [Dungeon&Fighter](https://df.nexon.com) user community QA board, extract 79k QA pairs data in 20k articles and 13k QA data using ChatGPT based on [DFU](https://www.dnf-universe.com/) articles
|
| 16 |
|
| 17 |
Fianl Model, bge-m3-ko_DNF is Mixed bge-m3 and trained model. Mixed reason detail read bge-m3 [Page](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune)
|
| 18 |
|
| 19 |
While the overall similarity has decreased, the discriminatory power between queries (questions) and chunks (answer search space) has increased, allowing for more accurate parsing of Dungeon & Fighters information.
|
| 20 |
|
|
|
|
| 21 |
<!--- Describe your model here -->
|
| 22 |
|
| 23 |
## Usage (Sentence-Transformers)
|