[XNNPACK] Serialize fp16 weights as fp16

Author: mcr229

Diff for: backends/xnnpack/operators/node_visitor.py

@@ -210,7 +210,7 @@ def get_serialized_dtype(
         self,
         quant_params: Optional[QuantParams],
         node: torch.fx.Node,
-        fp32_static_weight: bool = False,
+        force_fp32: bool = False,
     ) -> XNNDatatype:
         # Default initialization
         dtype = XNNDatatype.xnn_datatype_fp32
@@ -267,7 +267,7 @@ def get_per_channel_dtype(
             if node_dtype is not None and node_dtype == torch.float16:
                 dtype = (
                     XNNDatatype.xnn_datatype_fp32
-                    if fp32_static_weight
+                    if force_fp32
                     else XNNDatatype.xnn_datatype_fp16
                 )
 
@@ -348,7 +348,7 @@ def define_tensor(  # noqa: C901
         convert_to_nhwc: bool = False,
         swap_in_out_for_weights: bool = False,
         quant_params: Optional[QuantParams] = None,
-        fp32_static_weights: bool = False,
+        force_fp32: bool = False,
         groups: int = 1,
     ) -> None:
         """
@@ -368,7 +368,7 @@ def define_tensor(  # noqa: C901
                         constant data. If used along with convert_to_nhwc, this
                         swap will happen before converting to nhwc.
             quant_params: Quantization meta data for this tensor, None if it is not quantized
-            fp32_static_weights: XNN_FLAG_FP32_STATIC_WEIGHTS for fp16 conv
+            force_fp32: XNN_FLAG_force_fp32 for fp16 conv
             groups: number of groups for swap_in_out_for_weights
         """
 
@@ -405,7 +405,7 @@ def define_tensor(  # noqa: C901
             convert_to_nhwc,
             swap_in_out_for_weights,
             quant_params,
-            fp32_static_weights,
+            force_fp32,
             groups,
         )
 
@@ -418,7 +418,7 @@ def define_tensor(  # noqa: C901
             dims = [dims[i] for i in PERM_NCHW_TO_NHWC]
 
         dtype = self.get_serialized_dtype(
-            quant_params, tensor, fp32_static_weight=fp32_static_weights
+            quant_params, tensor, force_fp32=force_fp32
         )
 
         tvalue = XNNTensorValue(
@@ -504,7 +504,7 @@ def get_serialized_buffer_index(
         convert_to_nhwc: bool,
         swap_in_out_for_weights: bool,
         quant_params: Optional[QuantParams],
-        fp32_static_weights: bool = False,
+        force_fp32: bool = False,
         groups: int = 1,
     ) -> int:
         """
@@ -525,7 +525,7 @@ def get_serialized_buffer_index(
                         constant data. If used along with convert_to_nhwc, this
                         swap will happen before converting to nhwc.
             quant_params: Quantization meta data for this tensor, None if it is not quantize
-            fp32_static_weights: bool to indicate whether tensor is fp32 static weights
+            force_fp32: bool to indicate whether tensor is fp32 static weights
             groups: groups for swap_in_out_for_weights
 
         Returns:
@@ -554,7 +554,7 @@ def get_serialized_buffer_index(
         # Quantize buffer if static data is indeed quantized
         if quant_params is not None and not quant_params.is_dynamic:
             const_val = quant_params.quantize_tensor(const_val).contiguous()
-        elif const_val.dtype != torch.float16 or fp32_static_weights:
+        elif const_val.dtype != torch.float16 or force_fp32:
             # ensure that the const is fp32
             const_val = const_val.to(dtype=torch.float32).contiguous()
 

Diff for: backends/xnnpack/operators/op_conv2d.py

@@ -82,7 +82,6 @@ def define_node(
         weight_quant_params = QuantParams.from_weights(
             kernel_node, self._exported_program
         )
-        fp32_static_weights = kernel_node.meta["val"].dtype == torch.float16
 
         if weight_quant_params is not None and weight_quant_params.per_channel:
             if is_transpose:
@@ -102,7 +101,6 @@ def define_node(
             convert_to_nhwc=True,
             swap_in_out_for_weights=is_depthwise_conv or is_transpose,
             quant_params=weight_quant_params,
-            fp32_static_weights=fp32_static_weights,
             groups=groups if is_transpose else 1,
         )
         kwargs["filter_id"] = vals_to_ids[get_input_node(node, 1)]
@@ -127,13 +125,19 @@ def define_node(
             bias_quant_params = QuantParams.from_bias(
                 bias_node, weight_quant_params, input_quant_params
             )
+            # For dynamic quantization, there are no kernels with fp16 bias
+            # So we need to force the fp16 bias to fp32
+            force_fp32 = False
+            if input_quant_params is not None and input_quant_params.is_dynamic:
+                force_fp32 = True
+
             self.define_tensor(
                 get_input_node(node, 2),
                 xnn_graph,
                 vals_to_ids,
                 convert_to_nhwc=False,
                 quant_params=bias_quant_params,
-                fp32_static_weights=fp32_static_weights,
+                force_fp32=force_fp32,
             )
             kwargs["bias_id"] = vals_to_ids[get_input_node(node, 2)]
 

Diff for: backends/xnnpack/operators/op_linear.py

@@ -59,7 +59,6 @@ def define_node(
             xnn_graph,
             vals_to_ids,
             quant_params=weight_quant_params,
-            fp32_static_weights=True,
         )
         filter_id = vals_to_ids[weight_node]
 
@@ -69,12 +68,18 @@ def define_node(
             bias_quant_params = QuantParams.from_bias(
                 bias_node, weight_quant_params, input_quant_params
             )
+            # For dynamic quantization, there are no kernels with fp16 bias
+            # So we need to force the fp16 bias to fp32
+            force_fp32 = False
+            if input_quant_params is not None and input_quant_params.is_dynamic:
+                force_fp32 = True
+
             self.define_tensor(
                 get_input_node(node, 2),
                 xnn_graph,
                 vals_to_ids,
                 quant_params=bias_quant_params,
-                fp32_static_weights=True,
+                force_fp32=force_fp32,
             )
             bias_id = vals_to_ids[bias_node]
         else:

Diff for: backends/xnnpack/test/ops/test_linear.py

@@ -607,7 +607,7 @@ def _test_qd8_linear_per_tensor_unsupported(self, dtype: torch.dtype = torch.flo
                         tester.to_edge()
                         tester.partition(
                             Partition(DynamicallyQuantizedPartitioner)
-                        ).dump_artifact()
+                        )
                         # should have [add]mm node
                         if uses_bias:
                             tester.check(
@@ -624,7 +624,7 @@ def _test_qd8_linear_per_tensor_unsupported(self, dtype: torch.dtype = torch.flo
                     else:
                         tester.to_edge_transform_and_lower(
                             ToEdgeTransformAndLower([DynamicallyQuantizedPartitioner])
-                        ).dump_artifact()
+                        )
                         # should not have a delegate node
                         tester.check_not(
                             [
@@ -717,7 +717,7 @@ def test_fp16_linear(self):
                     num_batch_dims=num_batch_dims,
                     uses_bias=use_bias,
                     dtype=torch.float16,
-                    atol=5e-2,  # TODO(T212995726): Investigate right atol for rand[n] inputs
+                    atol=5e-3,  # TODO(T212995726): Investigate right atol for rand[n] inputs
                 )
 
     def test_fp32_linear(self):