Remove the warning message in V1 LSTM.

LSTM v2 is using v1 as base class, and warning message is misleading in that case.

When user starts using v2, the performance should be good since it will use cudnn kernel when available.

PiperOrigin-RevId: 238473081

tensorflow/python/keras/layers/recurrent.py

@@ -2508,10 +2508,6 @@ class LSTM(RNN):
       logging.warning('`implementation=0` has been deprecated, '
                       'and now defaults to `implementation=1`.'
                       'Please update your layer call.')
-    if context.executing_eagerly() and context.num_gpus() > 0:
-      logging.warn('%s: Note that this layer is not optimized for performance. '
-                   'Please use tf.keras.layers.CuDNNLSTM for better '
-                   'performance on GPU.', self)
     cell = LSTMCell(
         units,
         activation=activation,