add training details

README.md

@@ -18,12 +18,39 @@ This is a [sentence-transformers](https://www.SBERT.net) model:
 It maps sentences & paragraphs (text) into a 1024 dimensional dense vector space.
 The model is intended to be used together with [SetFit](https://github.com/huggingface/setfit)
 to improve German few-shot text classification.
+It has has a sibling model called
+[deutsche-telekom/gbert-large-paraphrase-cosine](https://huggingface.co/deutsche-telekom/gbert-large-paraphrase-cosine).
 
 This model is based on [deepset/gbert-large](https://huggingface.co/deepset/gbert-large).
 Many thanks to [deepset](https://www.deepset.ai/)!
 
 ## Training
-TODO
+
+**Loss Function**\
+We have used [BatchHardSoftMarginTripletLoss](https://www.sbert.net/docs/package_reference/losses.html#batchhardsoftmargintripletloss) with eucledian distance as the loss function:
+
+``` python
+    train_loss = losses.BatchHardSoftMarginTripletLoss(
+       model=model,
+       distance_metric=BatchHardTripletLossDistanceFunction.eucledian_distance,
+   )
+```
+
+**Training Data**\
+The model is trained on a carefully filtered dataset of
+[deutsche-telekom/ger-backtrans-paraphrase](https://huggingface.co/datasets/deutsche-telekom/ger-backtrans-paraphrase).
+We deleted the following pairs of sentences:
+- `min_char_len` less than 15
+- `jaccard_similarity` greater than 0.3
+- `de_token_count` greater than 30
+- `en_de_token_count` greater than 30
+- `cos_sim` less than 0.85
+
+**Hyperparameters**
+- learning_rate: 5.5512022294147105e-06
+- num_epochs: 7
+- train_batch_size: 68
+- num_gpu: ???
 
 ## Evaluation Results
 We use the [NLU Few-shot Benchmark - English and German](https://huggingface.co/datasets/deutsche-telekom/NLU-few-shot-benchmark-en-de)