Add generated files for unsuppported to DeformConv op

docs/Dialects/krnl.md

@@ -148,9 +148,9 @@ means to block the for loop referred to by %i using a tile size of 4.
 _Call operation_
 
 The call operation provides a generic way to replace an ONNX Op with a call
-to an external function at Krnl level. 
-`funcName` attributes determines which function to call. 
-`parameters` is the inputs to Krnl.Call. It includes the outputs and inputs 
+to an external function at Krnl level.
+`funcName` attributes determines which function to call.
+`parameters` is the inputs to Krnl.Call. It includes the outputs and inputs
 of the ONNX Op. The outputs and inputs are already lowered to MemRefs.
 The external function is assumed NOT to allocate or free any memory.
 'numOfOutput` attribute to tell how manu outputs Memref in parameters.

docs/Dialects/onnx.md

@@ -1941,6 +1941,46 @@ Effects: MemoryEffects::Effect{}
 | :----: | ----------- |
 | `output` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values or tensor of bfloat16 type values
 
+### `onnx.DeformConv` (ONNXDeformConvOp)
+
+_ONNX DeformConv operation_
+
+Performs deformable convolution as described in https://arxiv.org/abs/1703.06211 and https://arxiv.org/abs/1811.11168.
+This operator specification supports the general N-D case. Note that most common use cases have 2D or 3D data.
+
+Traits: AlwaysSpeculatableImplTrait
+
+Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface), ShapeHelperOpInterface, ShapeInferenceOpInterface
+
+Effects: MemoryEffects::Effect{}
+
+#### Attributes:
+
+| Attribute | MLIR Type | Description |
+| :-------: | :-------: | ----------- |
+| `dilations` | ::mlir::ArrayAttr | 64-bit integer array attribute
+| `group` | ::mlir::IntegerAttr | 64-bit signed integer attribute
+| `kernel_shape` | ::mlir::ArrayAttr | 64-bit integer array attribute
+| `offset_group` | ::mlir::IntegerAttr | 64-bit signed integer attribute
+| `pads` | ::mlir::ArrayAttr | 64-bit integer array attribute
+| `strides` | ::mlir::ArrayAttr | 64-bit integer array attribute
+
+#### Operands:
+
+| Operand | Description |
+| :-----: | ----------- |
+| `X` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values
+| `W` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values
+| `offset` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values
+| `B` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values or none type
+| `mask` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values or none type
+
+#### Results:
+
+| Result | Description |
+| :----: | ----------- |
+| `Y` | tensor of 16-bit float values or tensor of 32-bit float values or tensor of 64-bit float values
+
 ### `onnx.DepthToSpace` (ONNXDepthToSpaceOp)
 
 _ONNX DepthToSpace operation_
@@ -4042,15 +4082,15 @@ Effects: MemoryEffects::Effect{}
 
 _An operation that transforms data between different layout formats_
 
-An operation that transforms a tensor from a layout to another layout. 
+An operation that transforms a tensor from a layout to another layout.
 A layout is defined by an attribute, i.e. `target_layout`, which allows this
 operation work with an arbitrary layout (e.g. a layout used for accelerators).
 
 `target_layout` is optional. If it is not given, the input tensor will be
 transformed to a normal tensor that does not have layout.
 
 If `target_layout` is the same as the input's layout, this operation will
-become an no-op by canonicalization. 
+become an no-op by canonicalization.
 
 The input and output tensors must have the same shape.
 
@@ -4657,7 +4697,7 @@ MaxPool consumes an input tensor X and applies max pooling across
  ```
  pad_shape[i] = (output_spatial_shape[i] - 1) * strides_spatial_shape[i] + ((kernel_spatial_shape[i] - 1) * dilations[i] + 1) - input_spatial_shape[i]
  ```
- The output of each pooling window is maximum number of elements exclude pad. 
+ The output of each pooling window is maximum number of elements exclude pad.
 
 
 Traits: AlwaysSpeculatableImplTrait
@@ -8985,7 +9025,7 @@ _ONNX Softmax operation_
 
 The operator computes the normalized exponential values for the given input:
 
- Softmax(input, axis) = Exp(input) / ReduceSum(Exp(input), axis=axis, keepdims=1) 
+ Softmax(input, axis) = Exp(input) / ReduceSum(Exp(input), axis=axis, keepdims=1)
 
 The \"axis\" attribute indicates the dimension along which Softmax
 will be performed. The output tensor has the same shape

docs/Dialects/zhigh.md

@@ -87,14 +87,14 @@ Effects: MemoryEffects::Effect{}
 
 ZHigh 2D convolution operation
 
-ZHigh operation to perform 2D convolution. 
+ZHigh operation to perform 2D convolution.
 * input: `[num_batches, height_in, width_in, channels_in]`
-* input_kernel: `[kernel_height, kernel_width, channels_in, channels_out]` 
+* input_kernel: `[kernel_height, kernel_width, channels_in, channels_out]`
 * input_bias: `[channels_out] `
-* kernel_shape: 1D array of kernel height and width 
-* strides: 1D array of stride height and width 
-* padding_type: SAME_PADDING or VALID_PADDING 
-* act_func: ACT_NONE or ACT_RELU 
+* kernel_shape: 1D array of kernel height and width
+* strides: 1D array of stride height and width
+* padding_type: SAME_PADDING or VALID_PADDING
+* act_func: ACT_NONE or ACT_RELU
 * output: `[num_batches, height_out, width_out, channels_out]`
 
 Traits: AlwaysSpeculatableImplTrait
@@ -185,10 +185,10 @@ ZHigh GRU operation
 * Shape for input_weights is `[D, I, 3*H]`.
 * Shape for hidden_weights is `[D, H, 3*H]`.
 * Shape for input_bias and hidden_bias is `[D, 3*H]`.
-* Shape for hn_output is `[S, D, B, H]` if return all timesteps 
+* Shape for hn_output is `[S, D, B, H]` if return all timesteps
   and `[1, D, B, H]` if return the final step only.
-* S is timesteps, D is the number of directions (1 for unidirectional and 
-* 2 for bidirectional), B is batch size, I is input size, and 
+* S is timesteps, D is the number of directions (1 for unidirectional and
+* 2 for bidirectional), B is batch size, I is input size, and
 * H is hidden size.
 * direction accepts "forward", "reverse", or "bidirectional
 * return_all_steps: -1 returns all timesteps, 0: returns only the last timestep."
@@ -233,11 +233,11 @@ zHigh operation to perform a LSTM.
 * Shape for input_weights is  `[D, I, 4*H]`.
 * Shape for hidden_weights is  `[D, H, 4*H]`.
 * Shape for input_bias and hidden_bias is `[D, 4*H]`.
-* Shape for hn_output is `[S, D, B, H]` if return all timesteps 
+* Shape for hn_output is `[S, D, B, H]` if return all timesteps
   and `[1, D, B, H]` if return the final step only.
 * Shape for cf_output is `[1, D, B, H]`.
-* S is timesteps, D is the number of directions (1 for unidirectional and 
-* 2 for bidirectional), B is batch size, I is input size, and 
+* S is timesteps, D is the number of directions (1 for unidirectional and
+* 2 for bidirectional), B is batch size, I is input size, and
 * H is hidden size.
 * direction accepts "forward", "reverse", or "bidirectional
 * return_all_steps: -1 returns all timesteps, 0: returns only the last timestep.
@@ -387,8 +387,8 @@ Effects: MemoryEffects::Effect{}
 
 ZHigh 2D mean reduce operation
 
-ZHigh operation to perform 2D mean reduce. Given an input 4D tensor, 
-returns a downsampled tensor reducing the middle 2nd and 3rd dimensions 
+ZHigh operation to perform 2D mean reduce. Given an input 4D tensor,
+returns a downsampled tensor reducing the middle 2nd and 3rd dimensions
 to a size of 1 based on the mean of the original values.
  Input and Output tensors should be in the 3D layout.
 
@@ -546,7 +546,7 @@ Effects: MemoryEffects::Effect{}
 ZHigh stick operation for GRU
 
 ZHigh operation to perform a stick for GRU.
-Variadic: list of pointers for input data to be transformed: 
+Variadic: list of pointers for input data to be transformed:
   - GRU concatenated: 3 data pointers, one for each input gate in
 (Z)update, Reset, Hidden, (ZRH) gate order
 
@@ -575,9 +575,9 @@ Effects: MemoryEffects::Effect{}
 ZHigh stick operation for LSTM
 
 ZHigh operation to perform a stick for LSTM.
-Variadic: list of pointers for input data to be transformed: 
-  - LSTM concatenated: 4 data pointers, one for each input gate in 
-Forget, Input, Cell, Output (FICO) order, 
+Variadic: list of pointers for input data to be transformed:
+  - LSTM concatenated: 4 data pointers, one for each input gate in
+Forget, Input, Cell, Output (FICO) order,
 
 Traits: AlwaysSpeculatableImplTrait
 

docs/Dialects/zlow.md

@@ -525,7 +525,7 @@ Traits: MemRefsNormalizable
 ZLow stick operation for GRU
 
 ZLow operation to perform a stick for GRU.
-Variadic: list of pointers for input data to be transformed: 
+Variadic: list of pointers for input data to be transformed:
   - GRU concatenated: 3 data pointers, one for each input gate in (Z)update, Reset, Hidden, (ZRH) gate order.
 
 Traits: MemRefsNormalizable
@@ -550,7 +550,7 @@ Traits: MemRefsNormalizable
 ZLow stick operation for LSTM
 
 ZLow operation to perform a stick for LSTM.
-Variadic: list of pointers for input data to be transformed: 
+Variadic: list of pointers for input data to be transformed:
   - LSTM concatenated: 4 data pointers, one for each input gate in Forget, Input, Cell, Output (FICO) order.
 
 Traits: MemRefsNormalizable