NXP backend: Remove optimization in prune_cast_operators.py (#13377)

### Summary
Remove unnecessary optimization in prune_cast_operators.py in IR
optimizer.

### Test plan
All tests where a subgraph is delegated to Neutron.

cc @digantdesai @JakeStevens @robert-kalmar @Pop-Korn

Author: roman-janik-nxp

backends/nxp/backend/ir/tflite_optimizer/optimizations/prune_cast_operators.py

@@ -32,10 +32,6 @@
 from executorch.backends.nxp.backend.ir.tflite_optimizer.optimizations.permute_fully_connected_weights_after_reshape import (
     PermuteFullyConnectedWeightsAfterReshape,
 )
-from executorch.backends.nxp.backend.ir.tflite_optimizer.optimizations.prune_cast_operators import (
-    FuseCastOperators,
-    RemoveCastOperatorsWithNoEffect,
-)
 from executorch.backends.nxp.backend.ir.tflite_optimizer.optimizations.prune_quantize_operators import (
     FuseParallelQuantizeOperators,
     PruneQuantizeOperators,
@@ -71,9 +67,6 @@ class Optimization(Enum):
     ELIMINATE_DEAD_BRANCHES = 11
     PERMUTE_FULLY_CONNECTED_WEIGHTS_AFTER_RESHAPE = 12
 
-    FUSE_CAST_OPERATORS = 13
-    REMOVE_CAST_OPERATORS_WITH_NO_EFFECT = 14
-
     MOVE_ACTIVATION_BEFORE_CONCAT = 15
     COMBINE_HARD_SIGMOID_AND_MUL_INTO_HARD_SWISH = 16
 
@@ -141,12 +134,6 @@ def __init__(
             Optimization.PERMUTE_FULLY_CONNECTED_WEIGHTS_AFTER_RESHAPE: PermuteFullyConnectedWeightsAfterReshape(
                 builder, conversion_config
             ),
-            Optimization.FUSE_CAST_OPERATORS: FuseCastOperators(
-                builder, conversion_config
-            ),
-            Optimization.REMOVE_CAST_OPERATORS_WITH_NO_EFFECT: RemoveCastOperatorsWithNoEffect(
-                builder, conversion_config
-            ),
             Optimization.MOVE_ACTIVATION_BEFORE_CONCAT: MoveActivationBeforeConcatenation(
                 builder, conversion_config
             ),