add patent similarity model in text
Browse files
README.md
CHANGED
|
@@ -7,15 +7,24 @@ tags:
|
|
| 7 |
- feature-extraction
|
| 8 |
- sentence-similarity
|
| 9 |
- transformers
|
|
|
|
| 10 |
datasets:
|
| 11 |
- mpi-inno-comp/paecter_dataset
|
| 12 |
license: apache-2.0
|
| 13 |
---
|
| 14 |
|
| 15 |
-
# paecter
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
-
This is a [sentence-transformers](https://www.SBERT.net) model. This model is fine-tuned on patent texts, leveraging Google's BERT for Patents as its base.
|
| 18 |
-
It can be used to generate 1024 dimensional dense vector for patent texts for downstream tasks like semantic search, prior art search, clustering, and patent landscaping.
|
| 19 |
|
| 20 |
<!--- Describe your model here -->
|
| 21 |
|
|
|
|
| 7 |
- feature-extraction
|
| 8 |
- sentence-similarity
|
| 9 |
- transformers
|
| 10 |
+
- patent
|
| 11 |
datasets:
|
| 12 |
- mpi-inno-comp/paecter_dataset
|
| 13 |
license: apache-2.0
|
| 14 |
---
|
| 15 |
|
| 16 |
+
# paecter - a Patent Similarity Model
|
| 17 |
+
|
| 18 |
+
This model is a patent similarity model.
|
| 19 |
+
Built upon Google's BERT for Patents as its base model, it generates 1024-dimensional dense vector embeddings from patent text.
|
| 20 |
+
These vectors encapsulate the semantic essence of the given patent text, making it highly suitable for various downstream tasks related to patent analysis.
|
| 21 |
+
|
| 22 |
+
## Applications
|
| 23 |
+
* Semantic Search
|
| 24 |
+
* Prior Art Search
|
| 25 |
+
* Clustering
|
| 26 |
+
* Patent Landscaping
|
| 27 |
|
|
|
|
|
|
|
| 28 |
|
| 29 |
<!--- Describe your model here -->
|
| 30 |
|