diff --git a/params/params.py b/params/params.py
index 76ada4e..cfb126f 100644
--- a/params/params.py
+++ b/params/params.py
@@ -12,11 +12,11 @@ class Params:
     epochs = 300                         # training epochs
     batch_size = 52                      # batch size during training (is parallelization is True, each GPU has batch_size // num_gpus examples)
                                          # if using encoder_type 'convolutional' or 'generated', should be divisible by the number of languages 
-    learning_rate = 1e-3                 # starting learning rate
+    learning_rate = 8e-3                 # starting learning rate
     learning_rate_decay = 0.5            # decay multiplier used by step learning rate scheduler (use 0.5 for halving)
     learning_rate_decay_start = 15000    # number of training steps until the first lr decay, expected to be greater than learning_rate_decay_each
     learning_rate_decay_each = 15000     # size of the learning rate scheduler step in training steps, it decays lr every this number steps
-    learning_rate_encoder = 1e-3         # initial learning rate of the encoder, just used if encoder_optimizer is set to True
+    learning_rate_encoder = 8e-3         # initial learning rate of the encoder, just used if encoder_optimizer is set to True
     weight_decay = 1e-6                  # L2 regularization
     encoder_optimizer = False            # if True, different learning rates are used for the encoder and decoder, the ecoder uses learning_rate_encoder at start
     max_output_length = 5000             # maximal number of frames produced by decoder, the number of frames is usualy much lower during synthesis