NXP backend: Replace pass to fuse activations functions with joint quantization with activation (#14816)

### Summary
This PR replaces optimizations 'fuse_activation_functions.py' by
quantization of Conv 2D and Linear ops together with fusable activations
- selected activations supported by Neutron (Relu, Relu6, Sigmoid,
Tanh). Logic is determined by target specs, now supporting Neutron-C. Unified delegation workflow in `executorch_pipeline.py` and `aot_neutron_compile.py` and changed `NeutronEdgePassManager` usage.
Tests updated. Relu has now non-shared, standalone quantization. Small refactoring of unit tests.

### Test plan
Unit tests provided (test_edge_passes.py, test_quantizer.py).

cc @robert-kalmar @JakeStevens @digantdesai

Author: roman-janik-nxp

backends/nxp/backend/edge_helper.py

@@ -4,10 +4,32 @@
 # LICENSE file in the root directory of this source tree.
 
 import torch
+
+from executorch.exir.dialects._ops import ops as exir_ops
+
 from torch.fx import GraphModule, Node
 from torch.nn import Parameter
 
 
+QUANTIZE_OPERATORS = [
+    exir_ops.edge.quantized_decomposed.quantize_per_channel.default,
+    exir_ops.edge.quantized_decomposed.quantize_per_tensor.default,
+]
+
+DEQUANTIZE_OPERATORS = [
+    exir_ops.edge.quantized_decomposed.dequantize_per_channel.default,
+    exir_ops.edge.quantized_decomposed.dequantize_per_tensor.default,
+]
+
+
+def _is_dequantize(node_: Node) -> bool:
+    return node_.op == "call_function" and node_.target in DEQUANTIZE_OPERATORS
+
+
+def _is_quantize(node_: Node) -> bool:
+    return node_.op == "call_function" and node_.target in QUANTIZE_OPERATORS
+
+
 def input_tensor(node: Node, input_index: int) -> torch.Tensor:
     if len(node.all_input_nodes) <= input_index:
         raise IndexError
@@ -62,12 +84,6 @@ def node_is_effectively_static_tensor(
     if node_is_static_tensor(node, parameters_mapping):
         return True
 
-    def _is_dequantize(node_: Node) -> bool:
-        return node_.target.__name__ in {
-            "quantized_decomposed.dequantize_per_tensor.default",
-            "quantized_decomposed.dequantize_per_channel.default",
-        }
-
     return _is_dequantize(node) and node_is_static_tensor(
         node.args[0], parameters_mapping
     )

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

@@ -11,9 +11,6 @@
 
 from executorch.backends.nxp.backend.ir import logger
 from executorch.backends.nxp.backend.ir.conversion_config import ConversionConfig
-from executorch.backends.nxp.backend.ir.tflite_optimizer.optimizations.fuse_activation_functions import (
-    FuseActivationFunctions,
-)
 from executorch.backends.nxp.backend.ir.tflite_optimizer.optimizations.move_relu_before_concat import (
     MoveActivationBeforeConcatenation,
 )
@@ -27,8 +24,6 @@
 
 
 class Optimization(Enum):
-    FUSE_ACTIVATION_FUNCTIONS = 1
-
     FUSE_TRANSPOSE_OPERATORS = 5
     REMOVE_IDENTITY_TRANSPOSE_OPERATORS = 6
 
@@ -64,9 +59,6 @@ def __init__(
         self._builder = builder
 
         self.optimization_map = {
-            Optimization.FUSE_ACTIVATION_FUNCTIONS: FuseActivationFunctions(
-                builder, conversion_config
-            ),
             Optimization.FUSE_TRANSPOSE_OPERATORS: FuseTransposeOperators(
                 builder, conversion_config
             ),