Update training query links and add multilingual language tag
Browse files
README.md
CHANGED
|
@@ -4,6 +4,7 @@ base_model: BAAI/bge-m3
|
|
| 4 |
language:
|
| 5 |
- de
|
| 6 |
- en
|
|
|
|
| 7 |
library_name: sentence-transformers
|
| 8 |
pipeline_tag: sentence-similarity
|
| 9 |
tags:
|
|
@@ -54,9 +55,8 @@ The model is optimized for queries generated from IFC element metadata and maps
|
|
| 54 |
|
| 55 |
The v3 run used project-internal data artifacts and generated pair files.
|
| 56 |
|
| 57 |
-
- Query source:
|
| 58 |
-
- Expected mapping source:
|
| 59 |
-
- Pair data used for training: internal random-preselected JSONL artifact (project-internal)
|
| 60 |
- Hard-negative strategy: fallback mode with random_preselected selection, up to 2 hard negatives per record
|
| 61 |
|
| 62 |
Train/dev counts from run metadata:
|
|
|
|
| 4 |
language:
|
| 5 |
- de
|
| 6 |
- en
|
| 7 |
+
- multilingual
|
| 8 |
library_name: sentence-transformers
|
| 9 |
pipeline_tag: sentence-similarity
|
| 10 |
tags:
|
|
|
|
| 55 |
|
| 56 |
The v3 run used project-internal data artifacts and generated pair files.
|
| 57 |
|
| 58 |
+
- Query source files: [Training/query_generation/generated_queries](https://github.com/Hygros/ifc-kbob-ai-matcher/tree/main/Training/query_generation/generated_queries)
|
| 59 |
+
- Expected mapping source files: [Training/query_generation/generated_queries](https://github.com/Hygros/ifc-kbob-ai-matcher/tree/main/Training/query_generation/generated_queries)
|
|
|
|
| 60 |
- Hard-negative strategy: fallback mode with random_preselected selection, up to 2 hard negatives per record
|
| 61 |
|
| 62 |
Train/dev counts from run metadata:
|