Update README.md

README.md

@@ -29,8 +29,8 @@ The model has been fine-tuned on the German subsets of the following datasets:
 - [stsb_multi_mt](https://huggingface.co/datasets/stsb_multi_mt)
 
 The training procedure can be divided into two stages:
-- training on paraphrase corpora with the Multiple Negatives Ranking Loss
-- training on semantic textual similarity using theh Cosine Similarity Loss
+- training on paraphrase datasets with the Multiple Negatives Ranking Loss
+- training on semantic textual similarity datasets using the Cosine Similarity Loss
 
 # Results
 The model achieves the following results: