Update README.md

README.md

@@ -5,8 +5,26 @@ tags:
 - feature-extraction
 - sentence-similarity
 - transformers
-
+language:
+- de
 ---
+# {Overvie}
+Base-Model: gbert-base
+
+Fine-Tuning: sentence-transformer
+
+Training data: german sts-dataset (can be found [here](https://github.com/t-systems-on-site-services-gmbh/german-STSbenchmark))
+  * both aws und deepl machine translation are used
+  * Training on sts-train, sts-dev
+
+Evaluation data: german sts-dataset (sts-test)
+
+Infrastructure: GPU V100 (20GB)
+
+Hyperparameter:
+  * batch size 64
+  * epochs 4
+  * MultiNegativeRankingLoss
 
 # {MODEL_NAME}