Update README.md

README.md

@@ -32,13 +32,11 @@ The model can be used directly with a *question-answering* pipeline:
 ### Hyperparameters
 
 ```
-batch_size = na
+batch_size = 8
 n_epochs = 3
 base_LM_model = "ixambert-base-cased"
-max_seq_len = na
-learning_rate = na
-lr_schedule = na
-warmup_proportion = na
-doc_stride = na
-max_query_length= na
+learning_rate = 2e-5
+lr_schedule = "linear"
+max_seq_len = 384
+doc_stride = 128
 ```