Add support op and kernel (#2606)

* add op, test=develop

_all_pages/develop/support_operation_list.md

@@ -10,6 +10,7 @@ title: 支持OP列表
 - arg_max
 - assign
 - assign_value
+- attention_padding_mask
 - axpy
 - batch_norm
 - beam_search
@@ -20,13 +21,16 @@ title: 支持OP列表
 - calib
 - calib_once
 - cast
+- collect_fpn_proposals
 - concat
+- conditional_block
 - conv2d
 - conv2d_transpose
 - crop
 - decode_bboxes
 - density_prior_box
 - depthwise_conv2d
+- distribute_fpn_proposals
 - dropout
 - elementwise_add
 - elementwise_div
@@ -36,7 +40,9 @@ title: 支持OP列表
 - equal
 - exp
 - expand
+- fake_channel_wise_dequantize_max_abs
 - fake_dequantize_max_abs
+- fake_quantize_dequantize_moving_average_abs_max
 - fake_quantize_moving_average_abs_max
 - fake_quantize_range_abs_max
 - fc
@@ -52,6 +58,7 @@ title: 支持OP列表
 - fusion_elementwise_max_activation
 - fusion_elementwise_mul_activation
 - fusion_elementwise_sub_activation
+- gather
 - generate_proposals
 - graph_op
 - greater_equal
@@ -61,9 +68,11 @@ title: 支持OP列表
 - hard_sigmoid
 - im2sequence
 - increment
+- instance_norm
 - io_copy
 - io_copy_once
 - is_empty
+- layer_norm
 - layout
 - layout_once
 - leaky_relu
@@ -76,9 +85,12 @@ title: 支持OP列表
 - logical_or
 - logical_xor
 - lookup_table
+- lookup_table_v2
 - lrn
+- match_matrix_tensor
 - matmul
 - mean
+- merge_lod_tensor
 - mul
 - multiclass_nms
 - nearest_interp
@@ -94,23 +106,43 @@ title: 支持OP列表
 - read_from_array
 - reduce_max
 - reduce_mean
+- reduce_prod
+- reduce_sum
 - relu
 - relu6
 - relu_clipped
 - reshape
 - reshape2
 - roi_align
+- rsqrt
 - scale
+- search_aligned_mat_mul
+- search_attention_padding_mask
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_arithmetic
+- search_seq_depadding
+- search_seq_fc
+- search_seq_softmax
+- sequence_arithmetic
+- sequence_concat
 - sequence_expand
 - sequence_expand_as
 - sequence_pool
+- sequence_reshape
+- sequence_reverse
 - sequence_softmax
+- sequence_topk_avg_pooling
 - shape
 - shuffle_channel
 - sigmoid
 - slice
 - softmax
+- softsign
 - split
+- split_lod_tensor
+- sqrt
 - square
 - squeeze
 - squeeze2
@@ -123,6 +155,7 @@ title: 支持OP列表
 - uniform_random
 - unsqueeze
 - unsqueeze2
+- var_conv_2d
 - while
 - write_to_array
 - yolo_box
@@ -154,13 +187,16 @@ title: 支持OP列表
 - box_clip
 - box_coder
 - cast
+- collect_fpn_proposals
 - concat
+- conditional_block
 - conv2d
 - conv2d_transpose
 - crop
 - decode_bboxes
 - density_prior_box
 - depthwise_conv2d
+- distribute_fpn_proposals
 - dropout
 - elementwise_add
 - elementwise_div
@@ -172,12 +208,14 @@ title: 支持OP列表
 - expand
 - fc
 - fill_constant
+- fill_constant_batch_size_like
 - floor
 - fusion_elementwise_add_activation
 - fusion_elementwise_div_activation
 - fusion_elementwise_max_activation
 - fusion_elementwise_mul_activation
 - fusion_elementwise_sub_activation
+- gather
 - generate_proposals
 - greater_equal
 - greater_than
@@ -186,7 +224,11 @@ title: 支持OP列表
 - hard_sigmoid
 - im2sequence
 - increment
+- instance_norm
 - is_empty
+- layer_norm
+- layout
+- layout_once
 - leaky_relu
 - less_equal
 - less_than
@@ -197,8 +239,10 @@ title: 支持OP列表
 - logical_or
 - logical_xor
 - lookup_table
+- lookup_table_v2
 - lrn
 - matmul
+- merge_lod_tensor
 - mul
 - nearest_interp
 - negative
@@ -213,10 +257,12 @@ title: 支持OP列表
 - read_from_array
 - reduce_max
 - reduce_mean
+- reduce_prod
 - relu
 - relu6
 - relu_clipped
 - roi_align
+- rsqrt
 - scale
 - sequence_expand
 - sequence_pool
@@ -227,6 +273,7 @@ title: 支持OP列表
 - slice
 - softmax
 - split
+- split_lod_tensor
 - squeeze
 - squeeze2
 - stack
@@ -241,26 +288,97 @@ title: 支持OP列表
 - write_to_array
 - yolo_box
 
+
 ### X86 kernels
+- batch_norm
+- cast
 - concat
+- conv2d
+- depthwise_conv2d
+- dropout
 - elementwise_add
 - elementwise_sub
+- fc
 - fill_constant_batch_size_like
+- gather
+- gelu
 - gru
+- layer_norm
+- match_matrix_tensor
 - matmul
 - mul
+- pool2d
+- reduce_sum
 - relu
 - reshape
 - reshape2
 - scale
+- search_aligned_mat_mul
+- search_attention_padding_mask
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_arithmetic
+- search_seq_depadding
+- search_seq_fc
+- search_seq_softmax
+- sequence_arithmetic
+- sequence_concat
 - sequence_expand_as
 - sequence_pool
+- sequence_reverse
+- sequence_topk_avg_pooling
 - shape
 - slice
 - softmax
+- softsign
 - square
 - squeeze
 - squeeze2
+- stack
+- tanh
+- transpose
+- transpose2
+- var_conv_2d
+
+### CUDA kernels
+- attention_padding_mask
+- bilinear_interp
+- calib
+- concat
+- conv
+- dropout
+- elementwise_add
+- fusion_elementwise_add_activation
+- fusion_elementwise_mul_activation
+- elementwise_mul
+- feed
+- io_copy
+- layout
+- layout_once
+- leaky_relu
+- lookup_table
+- match_matrix_tensor
+- mul
+- nearest_interp
+- pool2d
+- relu
+- scale
+- search_aligned_mat_mul
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_depadding
+- search_seq_fc
+- sequence_arithmetic
+- sequence_concat
+- sequence_pool
+- sequence_reverse
+- sequence_topk_avg_pooling
+- softmax
+- transpose
+- var_conv_2d
+- yolo_box
 
 ### OpenCL kernels
 - conv2d
@@ -268,6 +386,8 @@ title: 支持OP列表
 - elementwise_add
 - fc
 - fusion_elementwise_add_activation
+- layout
+- layout_once
 - io_copy
 - io_copy_once
 - mul

_all_pages/v2.1.0/support_operation_list.md

@@ -10,6 +10,7 @@ title: 支持OP列表
 - arg_max
 - assign
 - assign_value
+- attention_padding_mask
 - axpy
 - batch_norm
 - beam_search
@@ -20,13 +21,16 @@ title: 支持OP列表
 - calib
 - calib_once
 - cast
+- collect_fpn_proposals
 - concat
+- conditional_block
 - conv2d
 - conv2d_transpose
 - crop
 - decode_bboxes
 - density_prior_box
 - depthwise_conv2d
+- distribute_fpn_proposals
 - dropout
 - elementwise_add
 - elementwise_div
@@ -36,7 +40,9 @@ title: 支持OP列表
 - equal
 - exp
 - expand
+- fake_channel_wise_dequantize_max_abs
 - fake_dequantize_max_abs
+- fake_quantize_dequantize_moving_average_abs_max
 - fake_quantize_moving_average_abs_max
 - fake_quantize_range_abs_max
 - fc
@@ -52,6 +58,7 @@ title: 支持OP列表
 - fusion_elementwise_max_activation
 - fusion_elementwise_mul_activation
 - fusion_elementwise_sub_activation
+- gather
 - generate_proposals
 - graph_op
 - greater_equal
@@ -61,9 +68,11 @@ title: 支持OP列表
 - hard_sigmoid
 - im2sequence
 - increment
+- instance_norm
 - io_copy
 - io_copy_once
 - is_empty
+- layer_norm
 - layout
 - layout_once
 - leaky_relu
@@ -76,9 +85,12 @@ title: 支持OP列表
 - logical_or
 - logical_xor
 - lookup_table
+- lookup_table_v2
 - lrn
+- match_matrix_tensor
 - matmul
 - mean
+- merge_lod_tensor
 - mul
 - multiclass_nms
 - nearest_interp
@@ -94,23 +106,43 @@ title: 支持OP列表
 - read_from_array
 - reduce_max
 - reduce_mean
+- reduce_prod
+- reduce_sum
 - relu
 - relu6
 - relu_clipped
 - reshape
 - reshape2
 - roi_align
+- rsqrt
 - scale
+- search_aligned_mat_mul
+- search_attention_padding_mask
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_arithmetic
+- search_seq_depadding
+- search_seq_fc
+- search_seq_softmax
+- sequence_arithmetic
+- sequence_concat
 - sequence_expand
 - sequence_expand_as
 - sequence_pool
+- sequence_reshape
+- sequence_reverse
 - sequence_softmax
+- sequence_topk_avg_pooling
 - shape
 - shuffle_channel
 - sigmoid
 - slice
 - softmax
+- softsign
 - split
+- split_lod_tensor
+- sqrt
 - square
 - squeeze
 - squeeze2
@@ -123,6 +155,7 @@ title: 支持OP列表
 - uniform_random
 - unsqueeze
 - unsqueeze2
+- var_conv_2d
 - while
 - write_to_array
 - yolo_box
@@ -154,13 +187,16 @@ title: 支持OP列表
 - box_clip
 - box_coder
 - cast
+- collect_fpn_proposals
 - concat
+- conditional_block
 - conv2d
 - conv2d_transpose
 - crop
 - decode_bboxes
 - density_prior_box
 - depthwise_conv2d
+- distribute_fpn_proposals
 - dropout
 - elementwise_add
 - elementwise_div
@@ -172,12 +208,14 @@ title: 支持OP列表
 - expand
 - fc
 - fill_constant
+- fill_constant_batch_size_like
 - floor
 - fusion_elementwise_add_activation
 - fusion_elementwise_div_activation
 - fusion_elementwise_max_activation
 - fusion_elementwise_mul_activation
 - fusion_elementwise_sub_activation
+- gather
 - generate_proposals
 - greater_equal
 - greater_than
@@ -186,7 +224,11 @@ title: 支持OP列表
 - hard_sigmoid
 - im2sequence
 - increment
+- instance_norm
 - is_empty
+- layer_norm
+- layout
+- layout_once
 - leaky_relu
 - less_equal
 - less_than
@@ -197,8 +239,10 @@ title: 支持OP列表
 - logical_or
 - logical_xor
 - lookup_table
+- lookup_table_v2
 - lrn
 - matmul
+- merge_lod_tensor
 - mul
 - nearest_interp
 - negative
@@ -213,10 +257,12 @@ title: 支持OP列表
 - read_from_array
 - reduce_max
 - reduce_mean
+- reduce_prod
 - relu
 - relu6
 - relu_clipped
 - roi_align
+- rsqrt
 - scale
 - sequence_expand
 - sequence_pool
@@ -227,6 +273,7 @@ title: 支持OP列表
 - slice
 - softmax
 - split
+- split_lod_tensor
 - squeeze
 - squeeze2
 - stack
@@ -241,26 +288,97 @@ title: 支持OP列表
 - write_to_array
 - yolo_box
 
+
 ### X86 kernels
+- batch_norm
+- cast
 - concat
+- conv2d
+- depthwise_conv2d
+- dropout
 - elementwise_add
 - elementwise_sub
+- fc
 - fill_constant_batch_size_like
+- gather
+- gelu
 - gru
+- layer_norm
+- match_matrix_tensor
 - matmul
 - mul
+- pool2d
+- reduce_sum
 - relu
 - reshape
 - reshape2
 - scale
+- search_aligned_mat_mul
+- search_attention_padding_mask
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_arithmetic
+- search_seq_depadding
+- search_seq_fc
+- search_seq_softmax
+- sequence_arithmetic
+- sequence_concat
 - sequence_expand_as
 - sequence_pool
+- sequence_reverse
+- sequence_topk_avg_pooling
 - shape
 - slice
 - softmax
+- softsign
 - square
 - squeeze
 - squeeze2
+- stack
+- tanh
+- transpose
+- transpose2
+- var_conv_2d
+
+### CUDA kernels
+- attention_padding_mask
+- bilinear_interp
+- calib
+- concat
+- conv
+- dropout
+- elementwise_add
+- fusion_elementwise_add_activation
+- fusion_elementwise_mul_activation
+- elementwise_mul
+- feed
+- io_copy
+- layout
+- layout_once
+- leaky_relu
+- lookup_table
+- match_matrix_tensor
+- mul
+- nearest_interp
+- pool2d
+- relu
+- scale
+- search_aligned_mat_mul
+- search_fc
+- search_grnn
+- search_group_padding
+- search_seq_depadding
+- search_seq_fc
+- sequence_arithmetic
+- sequence_concat
+- sequence_pool
+- sequence_reverse
+- sequence_topk_avg_pooling
+- softmax
+- transpose
+- var_conv_2d
+- yolo_box
 
 ### OpenCL kernels
 - conv2d
@@ -268,6 +386,8 @@ title: 支持OP列表
 - elementwise_add
 - fc
 - fusion_elementwise_add_activation
+- layout
+- layout_once
 - io_copy
 - io_copy_once
 - mul