legacy_backward.yaml 92.5 KB
Newer Older
1
- backward_op : abs_double_grad
Z
zyfncg 已提交
2 3 4 5 6 7 8 9 10 11 12
  forward : abs_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_x_grad)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : abs_double_grad
  data_transform:
    skip_transform : grad_x_grad

13
- backward_op : abs_grad
Z
zyfncg 已提交
14 15 16 17 18 19 20 21 22 23
  forward : abs (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : abs_grad
  backward : abs_double_grad

24
- backward_op : add_double_grad
Z
zyfncg 已提交
25 26 27 28 29 30 31 32 33 34 35 36
  forward : add_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
  args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [grad_out]
  kernel :
    func : add_double_grad
  optional : grad_x_grad, grad_y_grad
  backward : add_triple_grad
  inplace : (grad_x_grad -> grad_out_grad)

37
- backward_op : add_grad
Z
zyfncg 已提交
38 39 40 41 42 43 44 45 46 47 48 49
  forward : add (Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : add_grad
  no_need_buffer : x, y
  backward : add_double_grad
  inplace : (out_grad -> x_grad)

50
- backward_op : add_triple_grad
Z
zyfncg 已提交
51 52 53 54 55 56 57 58 59 60
  forward : add_double_grad (Tensor y, Tensor grad_out, Tensor grad_grad_x, Tensor grad_grad_y, int axis = -1) -> Tensor(grad_grad_out)
  args : (Tensor grad_grad_x, Tensor grad_grad_y, Tensor grad_grad_out_grad, int axis = -1)
  output : Tensor(grad_grad_x_grad), Tensor(grad_grad_y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [grad_grad_x, grad_grad_y]
  kernel :
    func : add_triple_grad
  inplace : (grad_grad_out_grad -> grad_grad_x_grad)

61
- backward_op : addmm_grad
62
  forward : addmm (Tensor input, Tensor x, Tensor y, float beta, float alpha) -> Tensor(out)
Z
zyfncg 已提交
63 64 65 66 67 68 69 70
  args : (Tensor input, Tensor x, Tensor y, Tensor out_grad, float alpha, float beta)
  output : Tensor(input_grad), Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [input, x, y]
  kernel :
    func : addmm_grad

71
- backward_op : affine_grid_grad
72 73 74 75 76 77 78 79 80 81 82
  forward : affine_grid (Tensor input, IntArray outputShape, bool use_cudnn=true, bool align_corners=true) -> Tensor(output)
  args : (Tensor output_grad, IntArray outputShape, bool use_cudnn=true, bool align_corners=true)
  output : Tensor(input_grad)
  infer_meta :
    func : AffineGridGradInferMeta
    param : [output_grad, outputShape, align_corners]
  kernel :
    func : affine_grid_grad
    param : [output_grad, outputShape, align_corners]
    use_gpudnn: use_cudnn

83
- backward_op : amax_grad
84 85
  forward: amax (Tensor x,  int64_t[] axis={},  bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis={},  bool keepdim=false, bool reduce_all=false)
86 87 88 89 90 91 92
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : amax_grad

93
- backward_op : amin_grad
94 95
  forward: amin (Tensor x,  int64_t[] axis={},  bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis={},  bool keepdim=false, bool reduce_all=false)
96 97 98 99 100 101 102
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : amin_grad

103
- backward_op : as_complex_grad
104 105 106 107 108
  forward : as_complex (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  invoke : as_real(out_grad)

109
- backward_op : as_real_grad
110 111 112 113 114
  forward : as_real (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  invoke : as_complex(out_grad)

115
- backward_op : assign_grad
Z
zyfncg 已提交
116 117 118
  forward : assign (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
119
  invoke : assign(out_grad)
Z
zyfncg 已提交
120

121
- backward_op : assign_out__grad
Z
zyfncg 已提交
122 123 124 125 126 127 128 129 130
  forward : assign_out_ (Tensor x, Tensor output) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
  kernel :
    func : assign
  inplace : (out_grad -> x_grad)

131
- backward_op : batch_norm_double_grad
Z
zyfncg 已提交
132 133 134 135 136 137 138 139 140 141 142 143
  forward : batch_norm_grad (Tensor x, Tensor scale, Tensor bias, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor grad_out, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
  args : (Tensor x, Tensor scale, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor grad_out,  Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x, scale, x]
  kernel :
    func : batch_norm_grad_grad
    data_type : x
  optional : out_mean, out_variance
  inplace : (grad_out -> grad_out_grad)

144
- backward_op : batch_norm_grad
Z
zyfncg 已提交
145 146 147 148 149 150 151 152 153 154 155 156
  forward : batch_norm (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
  args : (Tensor x, Tensor scale, Tensor bias, Tensor mean_out, Tensor variance_out, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor out_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x, scale, bias]
  kernel :
    func : batch_norm_grad
    data_type : out_grad
  optional : mean_out, variance_out, reserve_space
  backward : batch_norm_double_grad

157
- backward_op : bce_loss_grad
Z
zyfncg 已提交
158 159 160 161 162 163 164 165 166 167
  forward : bce_loss (Tensor input, Tensor label) -> Tensor(out)
  args : (Tensor input, Tensor label, Tensor out_grad)
  output : Tensor(input_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [input]
  kernel :
    func : bce_loss_grad
  inplace : (out_grad -> input_grad)

168
- backward_op : bicubic_interp_grad
169 170 171 172 173 174 175 176 177 178 179
  forward : bicubic_interp (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode) -> Tensor(output)
  args : (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, Tensor output_grad, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  optional: out_size, size_tensor, scale_tensor
  kernel :
    func : bicubic_interp_grad
    data_type : output_grad

180
- backward_op : bilinear_interp_grad
181 182 183 184 185 186 187 188 189 190 191
  forward : bilinear_interp (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode) -> Tensor(output)
  args : (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, Tensor output_grad, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  optional: out_size, size_tensor, scale_tensor
  kernel :
    func : bilinear_interp_grad
    data_type : output_grad

192
- backward_op : bilinear_tensor_product_grad
193 194 195 196 197 198 199 200
  forward : bilinear_tensor_product (Tensor x, Tensor y, Tensor weight, Tensor bias) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor weight, Tensor out_grad)
  output : Tensor(x_grad), Tensor(y_grad), Tensor(weight_grad), Tensor(bias_grad)
  infer_meta :
    func : BilinearTensorProductGradInferMeta
  kernel :
    func : bilinear_tensor_product_grad

201
- backward_op : broadcast_tensors_grad
202 203 204
  forward : broadcast_tensors (Tensor[] input) -> Tensor[](out)
  args : (Tensor[] input, Tensor[] out_grad)
  output : Tensor[](input_grad)
205 206
  infer_meta :
    func : UnchangedMultiInferMeta
207
    param : [input]
208 209 210
  kernel :
    func : broadcast_tensors_grad
    param : [out_grad]
211
  no_need_buffer : input
212

213
- backward_op : cast_grad
214
  forward : cast (Tensor x, DataType dtype) -> Tensor(out)
Z
zyfncg 已提交
215 216
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
217
  invoke : cast (out_grad, x.dtype())
Z
zyfncg 已提交
218 219
  no_need_buffer : x

220
- backward_op : ceil_grad
Z
zyfncg 已提交
221 222 223 224 225 226 227 228 229 230
  forward : ceil(Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [out_grad]
  kernel :
    func : ceil_grad
  inplace : (out_grad -> x_grad)

231
- backward_op : celu_double_grad
Z
zyfncg 已提交
232 233 234 235 236 237 238 239 240 241
  forward : celu_grad(Tensor x, Tensor grad_out, float alpha) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
  output : Tensor(x_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, x]
  kernel :
    func : celu_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

242
- backward_op : celu_grad
Z
zyfncg 已提交
243 244 245 246 247 248 249 250 251 252 253
  forward : celu(Tensor x, float alpha) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float alpha)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : celu_grad
  backward : celu_double_grad
  inplace : (out_grad -> x_grad)

254
- backward_op : clip_double_grad
Z
zyfncg 已提交
255 256 257 258 259 260 261 262 263
  forward : clip_grad (Tensor x, Tensor grad_out, Scalar min = 0., Scalar max = 0.) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_x_grad, Scalar min = 0., Scalar max = 0.)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : clip_grad

264
- backward_op : clip_grad
Z
zyfncg 已提交
265 266 267 268 269 270 271 272 273 274 275
  forward : clip (Tensor x, Scalar min, Scalar max) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, Scalar min = 0., Scalar max = 0.)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : clip_grad
  backward : clip_double_grad
  inplace : (out_grad -> x_grad)

276
- backward_op : complex_grad
277 278 279
  forward : complex (Tensor real, Tensor imag) -> Tensor(out)
  args : (Tensor real, Tensor imag, Tensor out_grad)
  output : Tensor(real_grad), Tensor(imag_grad)
280 281 282 283
  infer_meta :
    func : ComplexGradInferMeta
  kernel :
    func : complex_grad
284
    data_type : real
285

286
- backward_op : concat_double_grad
Z
zyfncg 已提交
287 288 289
  forward : concat_grad (Tensor[] x, Tensor grad_out, Scalar axis) -> Tensor[](grad_x)
  args : (Tensor[] grad_x_grad, Scalar axis = 0)
  output : Tensor(grad_out_grad)
290
  invoke : concat(grad_x_grad, axis)
Z
zyfncg 已提交
291

292
- backward_op : concat_grad
Z
zyfncg 已提交
293 294 295 296 297 298 299 300 301 302 303
  forward : concat (Tensor[] x, Scalar axis) -> Tensor(out)
  args : (Tensor[] x, Tensor out_grad, Scalar axis = 0)
  output : Tensor[](x_grad){x.size()}
  infer_meta :
    func : UnchangedMultiInferMeta
    param : [x]
  kernel :
    func : concat_grad
  no_need_buffer : x
  backward : concat_double_grad

304
- backward_op : conj_grad
Z
zyfncg 已提交
305 306 307 308 309 310 311 312 313
  forward : conj (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [out_grad]
  kernel :
    func : conj

314
- backward_op : conv2d_grad
Z
zyfncg 已提交
315 316 317
  forward : conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
  args : (Tensor input, Tensor filter, Tensor out_grad,  int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
  output : Tensor(input_grad), Tensor(filter_grad)
Z
zyfncg 已提交
318 319 320 321 322 323
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [input, filter]
  kernel :
    func : conv2d_grad
    use_gpudnn : true
Z
zyfncg 已提交
324 325
  backward : conv2d_grad_grad

326
- backward_op : conv2d_grad_grad
Z
zyfncg 已提交
327 328 329 330 331 332 333 334 335 336 337
  forward : conv2d_grad (Tensor input, Tensor filter, Tensor grad_out,  int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
  args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
  output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param: [input, filter, grad_out]
  kernel :
    func : conv2d_grad_grad
    use_gpudnn : true
  optional : grad_input_grad, grad_filter_grad

338
- backward_op : conv2d_transpose_double_grad
339 340
  forward : conv2d_transpose_grad(Tensor x, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(grad_x), Tensor(grad_filter)
  args : (Tensor x, Tensor filter, Tensor grad_out, Tensor grad_x_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
Z
zyfncg 已提交
341 342 343 344 345 346 347
  output : Tensor(x_grad), Tensor(filter_grad), Tensor(grad_out_grad)
  infer_meta :
    func : Conv2dTransposeDoubleGradInferMeta
  kernel :
    func : conv2d_transpose_grad_grad
    use_gpudnn : true

348
- backward_op : conv2d_transpose_grad
349 350
  forward : conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
  args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
Z
zyfncg 已提交
351 352
  output : Tensor(x_grad), Tensor(filter_grad)
  infer_meta :
353
    func : Conv2dTransposeGradInferMeta
Z
zyfncg 已提交
354 355 356 357 358
  kernel :
    func : conv2d_transpose_grad
    use_gpudnn : true
  backward : conv2d_transpose_double_grad

359
- backward_op : conv3d_grad
Z
zyfncg 已提交
360 361 362
  forward : conv3d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
  args : (Tensor input, Tensor filter, Tensor out_grad,  int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
  output : Tensor(input_grad), Tensor(filter_grad)
Z
zyfncg 已提交
363 364 365 366 367 368
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [input, filter]
  kernel :
    func : conv3d_grad
    use_gpudnn : true
Z
zyfncg 已提交
369 370
  backward : conv3d_grad_grad

371
- backward_op : conv3d_grad_grad
Z
zyfncg 已提交
372 373 374 375 376 377 378 379 380 381 382
  forward : conv3d_grad (Tensor input, Tensor filter, Tensor grad_out,  int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
  args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
  output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param: [input, filter, grad_out]
  kernel :
    func : conv3d_grad_grad
    use_gpudnn : true
  optional : grad_input_grad, grad_filter_grad

383
- backward_op : conv3d_transpose_grad
Z
zyfncg 已提交
384 385 386 387 388 389 390 391 392
  forward : conv3d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
  args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
  output : Tensor(x_grad), Tensor(filter_grad)
  infer_meta :
    func : ConvTransposeGradInferMeta
  kernel :
    func : conv3d_transpose_grad
    use_gpudnn : true

393
- backward_op : crop_grad
394 395 396 397
  forward : crop_tensor (Tensor x, IntArray shape, IntArray offsets) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray offsets)
  output : Tensor(x_grad)
  infer_meta :
398
    func : CropGradInferMeta
399
  kernel :
400
    func : crop_grad
401 402
    data_type : x

403
- backward_op : cross_entropy_with_softmax_grad
Z
zyfncg 已提交
404 405 406 407 408 409 410 411 412 413
  forward : cross_entropy_with_softmax (Tensor input, Tensor label, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis) -> Tensor(softmax), Tensor(loss)
  args : (Tensor label, Tensor softmax, Tensor loss_grad, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis)
  output : Tensor(input_grad)
  infer_meta :
    func : CrossEntropyWithSoftmaxGradInferMeta
  kernel :
    func : cross_entropy_with_softmax_grad
    data_type : softmax
  inplace : (softmax -> input_grad)

414
- backward_op : cumprod_grad
Z
zyfncg 已提交
415 416 417 418 419 420 421 422 423
  forward : cumprod (Tensor x, int dim) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int dim)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : cumprod_grad

424
- backward_op : cumsum_grad
W
WangZhen 已提交
425 426
  forward : cumsum(Tensor x, Scalar axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
  args : (Tensor out_grad, Scalar axis, bool flatten, bool exclusive, bool reverse)
Z
zyfncg 已提交
427 428 429
  output : Tensor(x_grad)
  invoke : cumsum(out_grad, axis, flatten, exclusive, !reverse)

430
- backward_op : deformable_conv_grad
Z
zyfncg 已提交
431 432 433 434 435 436 437 438 439 440
  forward : deformable_conv(Tensor x, Tensor offset, Tensor filter, Tensor mask, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step) -> Tensor(out)
  args : (Tensor x, Tensor offset, Tensor filter, Tensor mask, Tensor out_grad, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step)
  output : Tensor(x_grad), Tensor(offset_grad), Tensor(filter_grad), Tensor(mask_grad)
  infer_meta :
    func : DeformableConvGradInferMeta
  kernel :
    func : deformable_conv_grad
    data_type : x
  optional : mask

441
- backward_op : depthwise_conv2d_grad
Z
zyfncg 已提交
442 443 444 445 446 447 448 449 450 451 452 453
  forward : depthwise_conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(out)
  args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn)
  output : Tensor(input_grad), Tensor(filter_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [input, filter]
  kernel :
    func : depthwise_conv2d_grad
    param : [input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, fuse_relu]
    use_gpudnn : use_gpudnn
  backward : depthwise_conv2d_grad_grad

454
- backward_op : depthwise_conv2d_grad_grad
Z
zyfncg 已提交
455 456 457 458 459 460 461 462 463 464
  forward : depthwise_conv2d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(grad_input), Tensor(grad_filter)
  args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu)
  output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param: [input, filter, grad_out]
  kernel :
    func : depthwise_conv2d_grad_grad
  optional : grad_input_grad, grad_filter_grad

465
- backward_op : depthwise_conv2d_transpose_grad
466 467
  forward : depthwise_conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
  args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, IntArray output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
Z
zyfncg 已提交
468 469
  output : Tensor(x_grad), Tensor(filter_grad)
  infer_meta :
470
    func : Conv2dTransposeGradInferMeta
Z
zyfncg 已提交
471 472 473
  kernel :
    func : depthwise_conv2d_transpose_grad

474
- backward_op : divide_double_grad
Z
zyfncg 已提交
475 476 477 478 479 480 481 482 483 484 485 486
  forward : divide_grad (Tensor x, Tensor y, Tensor out, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
  args : (Tensor y, Tensor out, Tensor grad_x, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
  output : Tensor(y_grad), Tensor(out_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [y, grad_x, grad_x]
  kernel :
    func : divide_double_grad
    data_type : out
  optional : grad_x_grad, grad_y_grad
  inplace : (grad_x_grad -> grad_out_grad)

487
- backward_op : divide_grad
Z
zyfncg 已提交
488 489 490 491 492 493 494 495 496 497
  forward : divide (Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : divide_grad
  backward : divide_double_grad

498
- backward_op : dropout_grad
499 500
  forward : dropout (Tensor x, Tensor seed_tensor, Scalar p, bool is_test, str mode, int seed, bool fix_seed) -> Tensor(out), Tensor(mask)
  args : (Tensor mask, Tensor out_grad, Scalar p, bool is_test, str mode)
Z
zyfncg 已提交
501 502 503 504 505 506 507
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : dropout_grad

508
- backward_op : eig_grad
509 510 511 512 513 514 515 516 517 518 519 520
  forward : eig (Tensor x) -> Tensor(out_w), Tensor(out_v)
  args : (Tensor out_w, Tensor out_v, Tensor out_w_grad, Tensor out_v_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_v]
  kernel :
    func : eig_grad
    data_type : out_v
  data_transform:
    skip_transform : out_w, out_w_grad

521
- backward_op : eigh_grad
522
  forward : eigh (Tensor x, str UPLO) -> Tensor(out_w), Tensor(out_v)
Z
zyfncg 已提交
523 524 525 526 527 528 529 530 531 532 533
  args : (Tensor out_w, Tensor out_v, Tensor out_w_grad, Tensor out_v_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_v]
  kernel :
    func : eigh_grad
    data_type : out_v
  data_transform:
    skip_transform : out_w, out_w_grad

534
- backward_op : eigvalsh_grad
535 536 537 538 539 540 541 542 543 544 545
  forward : eigvalsh (Tensor x, str uplo, bool is_test) -> Tensor(eigenvalues), Tensor(eigenvectors)
  args : (Tensor eigenvectors, Tensor eigenvalues_grad, str uplo, bool is_test)
  output : Tensor(x_grad)
  infer_meta :
    func : EigvalshGradInferMeta
  kernel :
    func : eigvalsh_grad
    data_type : eigenvectors
  data_transform :
    skip_transform : eigenvalues_grad

546
- backward_op : einsum_grad
Z
zyfncg 已提交
547 548 549 550 551 552 553 554 555
  forward : einsum (Tensor[] x, str equation) -> Tensor(out), Tensor[](inner_cache), Tensor[](x_shape)
  args : (Tensor[] x_shape, Tensor[] inner_cache, Tensor out_grad, str equation)
  output : Tensor[](x_grad){x.size()}
  infer_meta :
    func : UnchangedMultiInferMeta
    param : [x_shape]
  kernel :
    func : einsum_grad

556
- backward_op : elementwise_pow_grad
Z
zyfncg 已提交
557 558 559 560 561 562 563 564 565
  forward : elementwise_pow(Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, y]
  kernel :
    func : elementwise_pow_grad

566
- backward_op : elu_double_grad
Z
zyfncg 已提交
567 568 569 570 571 572 573 574 575 576
  forward : elu_grad (Tensor x, Tensor out, Tensor grad_out, float alpha)-> Tensor(grad_x)
  args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
  output : Tensor(x_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, x]
  kernel :
    func : elu_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

577
- backward_op : elu_grad
Z
zyfncg 已提交
578 579 580 581 582 583 584 585 586 587 588
  forward : elu (Tensor x, float alpha) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, float alpha)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : elu_grad
  backward : elu_double_grad
  inplace : (out_grad -> x_grad)

589
- backward_op : embedding_grad
Z
zyfncg 已提交
590 591 592 593 594
  forward : embedding (Tensor x, Tensor weight, int64_t padding_idx=-1, bool sparse=false) -> Tensor(out)
  args : (Tensor x, Tensor weight, Tensor out_grad, int64_t padding_idx=-1, bool sparse=false)
  output : Tensor(weight_grad)
  invoke : embedding_grad_impl(x, weight, out_grad, padding_idx, sparse, weight_grad)

595
- backward_op : expand_as_grad
Z
zyfncg 已提交
596 597 598 599 600 601 602 603 604 605
  forward : expand_as (Tensor x, Tensor y, int[] target_shape) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int[] target_shape)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : expand_as_grad
  no_need_buffer : x

606
- backward_op : expand_double_grad
Z
zyfncg 已提交
607 608 609
  forward : expand_grad (Tensor x, Tensor grad_out, IntArray shape) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray shape)
  output : Tensor(grad_out_grad)
610
  invoke : expand(grad_x_grad, shape)
Z
zyfncg 已提交
611

612
- backward_op : expand_grad
Z
zyfncg 已提交
613 614 615 616 617 618 619 620 621 622 623
  forward : expand (Tensor x, IntArray shape) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray shape)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : expand_grad
  no_need_buffer : x
  backward : expand_double_grad

624
- backward_op : expm1_grad
Z
zyfncg 已提交
625 626 627 628 629 630 631 632 633 634
  forward : expm1 (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : expm1_grad
  inplace : (out_grad -> x_grad)

635
- backward_op : exponential__grad
636
  forward : exponential_ (Tensor x, float lam) -> Tensor(out)
637 638 639 640
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
641
  invoke : zeros_like(out_grad)
642

643
- backward_op : fill_diagonal_grad
Z
zhiboniu 已提交
644 645 646 647 648 649 650
  forward : fill_diagonal (Tensor x, float value, int offset, bool wrap) -> Tensor(out)
  args : (Tensor out_grad, float value, int offset, bool wrap)
  output : Tensor(x_grad)
  infer_meta :
    func : FillDiagonalGradInferMeta
  kernel :
    func : fill_diagonal_grad
Z
zhiboniu 已提交
651

652
- backward_op : fill_diagonal_tensor_grad
Z
zhiboniu 已提交
653 654 655 656 657 658 659
  forward : fill_diagonal_tensor (Tensor x, Tensor y, int64_t offset, int dim1, int dim2) -> Tensor(out)
  args : (Tensor out_grad, int64_t offset, int dim1, int dim2)
  output : Tensor(x_grad)
  infer_meta :
    func : FillDiagonalTensorGradInferMeta
  kernel :
    func : fill_diagonal_tensor_grad
660 661
  inplace : (out_grad -> x_grad)

662
- backward_op : fill_grad
663 664 665 666 667 668 669 670 671 672
  forward : fill (Tensor x, Scalar value) -> Tensor(out)
  args : (Tensor out_grad, Scalar value)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : fill_grad
  inplace : (out_grad -> x_grad)

673
- backward_op : flatten_grad
Z
zyfncg 已提交
674 675 676 677 678 679 680 681 682 683 684 685 686
  forward : flatten(Tensor x, int start_axis, int stop_axis) -> Tensor(out), Tensor(xshape)
  args : (Tensor xshape, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func :  KernelWithXShapeInferMeta
    param : [xshape]
  kernel :
    func : flatten_grad
    data_type: out_grad
    backend: out_grad
    layout: out_grad
  inplace : (out_grad -> x_grad)

687
- backward_op : floor_grad
Z
zyfncg 已提交
688 689 690 691 692 693 694 695 696 697
  forward : floor(Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [out_grad]
  kernel :
    func : floor_grad
  inplace : (out_grad -> x_grad)

698
- backward_op : fmax_grad
Z
zyfncg 已提交
699 700 701 702 703 704 705 706 707
  forward : fmax(Tensor x, Tensor y, int axis) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, y]
  kernel :
    func : fmax_grad

708
- backward_op : fmin_grad
Z
zyfncg 已提交
709 710 711 712 713 714 715 716 717
  forward : fmin(Tensor x, Tensor y, int axis) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, y]
  kernel :
    func : fmin_grad

718
- backward_op : frame_grad
C
Charles-hit 已提交
719 720 721 722 723 724 725 726 727
  forward : frame(Tensor x, int frame_length, int hop_length, int axis) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int frame_length, int hop_length, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : frame_grad

728
- backward_op : frobenius_norm_grad
Z
zyfncg 已提交
729 730 731 732 733 734 735 736 737
  forward : frobenius_norm(Tensor x, int64_t[] axis,  bool keep_dim,  bool reduce_all) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis,  bool keep_dim,  bool reduce_all)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : frobenius_norm_grad

738
- backward_op : gather_grad
Z
zyfncg 已提交
739 740 741 742 743 744 745 746 747 748 749
  forward : gather(Tensor x, Tensor index, Scalar axis=0) -> Tensor(out)
  args : (Tensor x, Tensor index, Tensor out_grad, Scalar axis=0, bool overwrite=false)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    data_type: x
    func : gather_grad
  no_need_buffer : x

750
- backward_op : gather_nd_grad
Z
zyfncg 已提交
751 752 753 754 755 756 757 758 759 760
  forward : gather_nd (Tensor x, Tensor index) -> Tensor(out)
  args : (Tensor x, Tensor index, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : gather_nd_grad
  no_need_buffer : x

761
- backward_op : gelu_grad
Z
zyfncg 已提交
762 763 764 765 766 767 768 769 770
  forward : gelu(Tensor x,  bool approximate) -> Tensor(out)
  args : (Tensor x, Tensor out_grad,  bool approximate)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : gelu_grad

771
- backward_op : grid_sample_grad
W
Wang Bojun 已提交
772 773 774
  forward : grid_sample (Tensor x, Tensor grid, str mode, str padding_mode, bool align_corners) -> Tensor(out)
  args : (Tensor x, Tensor grid, Tensor out_grad, str mode, str padding_mode, bool align_corners)
  output : Tensor(x_grad), Tensor(grid_grad)
775
  infer_meta :
W
Wang Bojun 已提交
776 777
    func : GeneralBinaryGradInferMeta
    param : [x, grid]
778
  kernel :
W
Wang Bojun 已提交
779 780 781
    func : grid_sample_grad
    data_type : x

782
- backward_op : group_norm_grad
Z
zyfncg 已提交
783 784 785 786 787 788 789 790 791 792 793 794
  forward : group_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout) -> Tensor(y), Tensor(mean), Tensor(variance)
  args : (Tensor x, Tensor scale, Tensor bias, Tensor y, Tensor mean, Tensor variance, Tensor y_grad, float epsilon, int groups, str data_layout)
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [y, scale, bias]
  kernel :
    func : group_norm_grad
    data_type : y_grad
  optional: scale, bias
  inplace : (y_grad -> x_grad)

795
- backward_op : gumbel_softmax_grad
Z
zyfncg 已提交
796 797 798 799 800 801 802 803 804
  forward : gumbel_softmax (Tensor x, float temperature, bool hard, int axis) -> Tensor(out)
  args : (Tensor out, Tensor out_grad, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : GumbelSoftmaxGradInferMeta
    param : [out, out_grad, axis]
  kernel :
    func : gumbel_softmax_grad

805 806
- backward_op : hardshrink_grad
  forward : hardshrink (Tensor x, float threshold) -> Tensor(out)
Z
zyfncg 已提交
807 808 809 810 811 812 813 814 815
  args : (Tensor x, Tensor out_grad, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : hard_shrink_grad
  inplace : (out_grad -> x_grad)

816 817
- backward_op : hardsigmoid_grad
  forward : hardsigmoid (Tensor x, float slope, float offset) -> Tensor(out)
Z
zyfncg 已提交
818 819 820 821 822 823 824 825 826
  args : (Tensor out, Tensor out_grad, float slope, float offset)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : hard_sigmoid_grad
  inplace : (out_grad -> x_grad)

827 828
- backward_op : hardswish_grad
  forward : hardswish (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0) -> Tensor(out)
Z
zyfncg 已提交
829 830 831 832 833 834 835 836 837
  args : (Tensor x, Tensor out_grad, float threshold, float scale, float offset)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : hard_swish_grad
  inplace : (out_grad -> x_grad)

838 839 840 841 842 843 844 845 846 847 848
- backward_op : hardtanh_grad
  forward : hardtanh (Tensor x, float t_min, float t_max) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float t_min, float t_max)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : hard_tanh_grad
  inplace : (out_grad -> x_grad)

849
- backward_op : hierarchical_sigmoid_grad
850 851 852 853 854 855 856 857 858 859
  forward : hierarchical_sigmoid (Tensor x, Tensor w, Tensor label, Tensor path, Tensor code, Tensor bias, int num_classes, bool remote_prefetch, int trainer_id, int64_t[] height_sections, str[] epmap, str[] table_names, bool is_sparse) -> Tensor(out), Tensor(pre_out), Tensor(w_out)
  args : (Tensor x, Tensor w, Tensor label, Tensor path, Tensor code, Tensor bias, Tensor pre_out, Tensor out_grad, int num_classes, bool remote_prefetch, int trainer_id, int64_t[] height_sections, str[] epmap, str[] table_names, bool is_sparse)
  output : Tensor(x_grad), Tensor(w_grad), Tensor(bias_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x ,w, bias]
  optional: path, code, bias
  kernel :
    func : hierarchical_sigmoid_grad

860
- backward_op : huber_loss_grad
Z
zyfncg 已提交
861 862 863 864 865 866 867 868 869
  forward : huber_loss (Tensor input, Tensor label, float delta) -> Tensor(out), Tensor(residual)
  args : (Tensor residual, Tensor out_grad, float delta)
  output : Tensor(input_grad), Tensor(label_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [residual, residual]
  kernel :
    func : huber_loss_grad

870
- backward_op : imag_grad
Z
zyfncg 已提交
871 872 873 874 875
  forward : imag (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  invoke : imag_grad_impl(out_grad, x_grad)

876
- backward_op : index_add_grad
L
Li Min 已提交
877 878 879 880 881 882 883 884 885 886
  forward : index_add(Tensor x, Tensor index,  Tensor add_value, int axis) -> Tensor(out)
  args : (Tensor index, Tensor add_value, Tensor out_grad, int axis)
  output : Tensor(x_grad), Tensor(add_value_grad)
  infer_meta :
    func : IndexAddGradInferMeta
  kernel :
    func : index_add_grad
    data_type : out_grad
  inplace : (out_grad -> x_grad)

887
- backward_op : index_sample_grad
Z
zyfncg 已提交
888 889 890 891 892 893 894 895 896 897 898
  forward : index_sample (Tensor x, Tensor index) -> Tensor(out)
  args : (Tensor x, Tensor index, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : index_sample_grad
    data_type : out_grad
  no_need_buffer : x

899
- backward_op : index_select_grad
900 901
  forward : index_select(Tensor x, Tensor index,  int axis) -> Tensor(out)
  args : (Tensor x, Tensor index, Tensor out_grad,  int axis)
Z
zyfncg 已提交
902 903 904 905 906 907 908 909 910
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : index_select_grad
    data_type : x
  no_need_buffer : x

911
- backward_op : instance_norm_double_grad
Z
zyfncg 已提交
912 913 914 915 916 917 918 919 920 921
  forward : instance_norm_grad(Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, float epsilon) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
  args : (Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float epsilon)
  output : Tensor(x_grad), Tensor(fwd_scale_grad), Tensor(grad_y_grad)
  infer_meta :
    func : InstanceNormDoubleGradInferMeta
  kernel :
    func : instance_norm_double_grad
    data_type : x
  optional : fwd_scale, grad_x_grad, grad_scale_grad, grad_bias_grad

922
- backward_op : instance_norm_grad
Z
zyfncg 已提交
923 924 925 926 927 928 929 930 931 932 933
  forward : instance_norm(Tensor x, Tensor scale, Tensor bias, float epsilon) -> Tensor(y), Tensor(saved_mean), Tensor(saved_variance)
  args : (Tensor x, Tensor scale, Tensor saved_mean, Tensor saved_variance, Tensor y_grad, float epsilon)
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
  infer_meta :
    func : InstanceNormGradInferMeta
  kernel :
    func : instance_norm_grad
    data_type : x
  optional : scale
  backward : instance_norm_double_grad

934
- backward_op : inverse_grad
935 936 937 938 939 940 941 942
  forward : inverse(Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta:
    func : InverseGradInferMeta
  kernel :
    func : inverse_grad

943
- backward_op : kldiv_loss_grad
Z
zyfncg 已提交
944 945 946 947 948 949 950 951 952 953
  forward : kldiv_loss(Tensor x, Tensor label, str reduction) -> Tensor(out)
  args : (Tensor x, Tensor label, Tensor out_grad, str reduction)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : kldiv_loss_grad
  no_need_buffer : x

954
- backward_op : kron_grad
Z
zyfncg 已提交
955 956 957 958 959 960 961 962 963 964
  forward : kron (Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : kron_grad
    data_type : out_grad

965
- backward_op : kthvalue_grad
Z
zyfncg 已提交
966 967 968 969 970 971 972 973 974
  forward : kthvalue(Tensor x, int k, int axis, bool keepdim) -> Tensor(out), Tensor(indices)
  args : (Tensor x, Tensor indices, Tensor out_grad, int k, int axis, bool keepdim)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : kthvalue_grad

975
- backward_op : label_smooth_grad
Z
zyfncg 已提交
976 977 978 979 980 981 982 983 984
  forward : label_smooth (Tensor label, Tensor prior_dist, float epsilon) -> Tensor(out)
  args : (Tensor out_grad, float epsilon)
  output : Tensor(label_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : label_smooth_grad

985
- backward_op : layer_norm_grad
Z
zyfncg 已提交
986 987 988 989 990 991 992 993 994 995 996 997
  forward : layer_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int begin_norm_axis, bool is_test) -> Tensor(out), Tensor(mean), Tensor(variance)
  args : (Tensor x,  Tensor scale, Tensor bias, Tensor mean, Tensor variance, Tensor out_grad, float epsilon, int begin_norm_axis, bool is_test)
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
  infer_meta :
    func : LayerNormGradInferMeta
    param : [x, scale, bias]
  kernel :
    func : layer_norm_grad
    data_type : out_grad
  no_need_buffer : bias
  optional : scale, bias

998
- backward_op : leaky_relu_double_grad
999 1000
  forward : leaky_relu_grad (Tensor x, Tensor grad_out, float negative_slope) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_x_grad, float negative_slope)
Z
zyfncg 已提交
1001 1002 1003 1004 1005 1006 1007 1008
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [grad_x_grad]
  kernel :
    func : leaky_relu_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

1009
- backward_op : leaky_relu_grad
1010 1011
  forward : leaky_relu (Tensor x, float negative_slope) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float negative_slope)
Z
zyfncg 已提交
1012 1013 1014 1015 1016 1017 1018 1019 1020
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : leaky_relu_grad
  backward : leaky_relu_double_grad
  inplace : (out_grad -> x_grad)

1021
- backward_op : lerp_grad
Z
zyfncg 已提交
1022 1023 1024 1025 1026 1027 1028 1029 1030
  forward : lerp (Tensor x, Tensor y, Tensor weight) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor weight, Tensor out, Tensor out_grad)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : lerp_grad

1031
- backward_op : linear_interp_grad
1032
  forward : linear_interp (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode) -> Tensor(output)
1033 1034 1035 1036 1037 1038 1039
  args : (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, Tensor output_grad, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  optional: out_size, size_tensor, scale_tensor
  kernel :
1040
    func : linear_interp_grad
1041 1042
    data_type : output_grad

1043
- backward_op : log10_grad
Z
zyfncg 已提交
1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
  forward : log10 (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : log10_grad
  inplace : (out_grad -> x_grad)

1054
- backward_op : log1p_grad
Z
zyfncg 已提交
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
  forward : log1p (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : log1p_grad
  inplace : (out_grad -> x_grad)

1065
- backward_op : log2_grad
Z
zyfncg 已提交
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
  forward : log2 (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : log2_grad
  inplace : (out_grad -> x_grad)

1076
- backward_op : log_double_grad
Z
zyfncg 已提交
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
  forward : log_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
  output : Tensor(x_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, x]
  kernel :
    func : log_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

1087
- backward_op : log_grad
Z
zyfncg 已提交
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098
  forward : log (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : log_grad
  backward : log_double_grad
  inplace : (out_grad -> x_grad)

1099
- backward_op : log_loss_grad
Z
zyfncg 已提交
1100 1101 1102 1103 1104 1105 1106 1107 1108
  forward : log_loss (Tensor input, Tensor label, float epsilon) -> Tensor(out)
  args : (Tensor input, Tensor label, Tensor out_grad, float epsilon)
  output : Tensor(input_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [input]
  kernel :
    func : log_loss_grad

1109
- backward_op : log_softmax_grad
Z
zyfncg 已提交
1110 1111 1112 1113 1114 1115 1116 1117 1118
  forward : log_softmax(Tensor x,  int axis) -> Tensor(out)
  args : (Tensor out, Tensor out_grad,  int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [out]
  kernel :
    func : log_softmax_grad

1119
- backward_op : logcumsumexp_grad
Z
zyfncg 已提交
1120 1121 1122 1123 1124 1125 1126 1127 1128
  forward : logcumsumexp(Tensor x, int axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  args : (Tensor x, Tensor out, Tensor out_grad, int axis, bool flatten, bool exclusive, bool reverse)
  output : Tensor(x_grad)
  kernel :
    func : logcumsumexp_grad

1129
- backward_op : logit_grad
Z
zyfncg 已提交
1130 1131 1132 1133 1134 1135 1136 1137 1138
  forward : logit (Tensor x, float eps = 1e-6f) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float eps)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : logit_grad

1139
- backward_op : logsigmoid_grad
Z
zyfncg 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
  forward : logsigmoid (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : logsigmoid_grad
  inplace : (out_grad -> x_grad)

1150
- backward_op : logsumexp_grad
Z
zyfncg 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159
  forward : logsumexp(Tensor x, int64_t[] axis,  bool keepdim,  bool reduce_all) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis,  bool keepdim,  bool reduce_all)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : logsumexp_grad

1160
- backward_op : lu_grad
L
Lin Manhui 已提交
1161 1162 1163 1164 1165 1166 1167 1168
  forward : lu (Tensor x, bool pivot) -> Tensor(out), Tensor(pivots), Tensor(infos)
  args : (Tensor x, Tensor out, Tensor pivots, Tensor out_grad, bool pivot)
  output : Tensor(x_grad)
  infer_meta :
    func : LUGradInferMeta
  kernel :
    func : lu_grad

1169
- backward_op : lu_unpack_grad
1170 1171
  forward : lu_unpack (Tensor x, Tensor y, bool unpack_ludata, bool unpack_pivots) -> Tensor(pmat), Tensor(l), Tensor(u)
  args : (Tensor x, Tensor y, Tensor l, Tensor u, Tensor pmat, Tensor l_grad, Tensor u_grad, bool unpack_ludata, bool unpack_pivots)
1172 1173 1174 1175 1176 1177
  output : Tensor(x_grad)
  infer_meta :
    func : LUUnpackGradInferMeta
  kernel :
    func : lu_unpack_grad

1178
- backward_op : margin_cross_entropy_grad
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
  forward : margin_cross_entropy (Tensor logits, Tensor label, bool return_softmax, int ring_id, int rank, int nranks, float margin1, float margin2, float margin3, float scale) -> Tensor(softmax), Tensor(loss)
  args : (Tensor logits, Tensor label, Tensor softmax, Tensor loss_grad, bool return_softmax, int ring_id, int rank, int nranks, float margin1, float margin2, float margin3, float scale)
  output : Tensor(logits_grad)
  infer_meta :
    func : MarginCrossEntropyGradInferMeta
  kernel :
    func : margin_cross_entropy_grad
    data_type : softmax
  inplace : (softmax -> logits_grad)

1189
- backward_op : masked_select_grad
Z
zyfncg 已提交
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
  forward : masked_select (Tensor x, Tensor mask) -> Tensor(out)
  args : (Tensor x, Tensor mask, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : masked_select_grad
    data_type : x
  no_need_buffer : x

1201
- backward_op : matmul_double_grad
Z
zyfncg 已提交
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
  forward : matmul_grad (Tensor x, Tensor y, Tensor grad_out, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y)
  args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, bool transpose_x=false, bool transpose_y=false)
  output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x, y, grad_out]
  kernel :
    func : matmul_double_grad
  backward : matmul_triple_grad
  optional : grad_x_grad, grad_y_grad

1213
- backward_op : matmul_grad
Z
zyfncg 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
  forward : matmul (Tensor x, Tensor y, bool transpose_x=false, bool transpose_y=false) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, bool transpose_x=false, bool transpose_y=false)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : matmul_grad
  backward : matmul_double_grad

1224
- backward_op : matmul_triple_grad
Z
zyfncg 已提交
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
  forward : matmul_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
  args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, bool transpose_x=false, bool transpose_y=false)
  output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
  infer_meta :
    func : GeneralQuinaryGradInferMeta
    param : [x, y, fwd_grad_out, fwd_grad_grad_x, fwd_grad_grad_y]
  kernel :
    func : matmul_triple_grad
  optional : grad_x_grad, grad_y_grad, grad_grad_out_grad

1235
- backward_op : matrix_power_grad
Z
zyfncg 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244
  forward : matrix_power (Tensor x, int n) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int n)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : matrix_power_grad

1245
- backward_op : max_grad
1246 1247
  forward: max (Tensor x,  IntArray axis={},  bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, IntArray axis={}, bool keepdim=false, bool reduce_all=false)
Z
zyfncg 已提交
1248 1249 1250 1251 1252 1253 1254
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : max_grad

1255
- backward_op : max_pool2d_with_index_grad
Z
zyfncg 已提交
1256 1257 1258 1259 1260 1261 1262 1263
  forward : max_pool2d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
  args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
  output : Tensor(x_grad)
  infer_meta :
    func : MaxPoolWithIndexGradInferMeta
  kernel :
    func : max_pool2d_with_index_grad

1264
- backward_op : max_pool3d_with_index_grad
Z
zyfncg 已提交
1265 1266 1267 1268 1269 1270 1271 1272
  forward : max_pool3d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
  args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
  output : Tensor(x_grad)
  infer_meta :
    func : MaxPoolWithIndexGradInferMeta
  kernel :
    func : max_pool3d_with_index_grad

1273
- backward_op : maximum_grad
Z
zyfncg 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282
  forward : maximum(Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, y]
  kernel :
    func : maximum_grad

1283
- backward_op : maxout_grad
Z
zyfncg 已提交
1284 1285 1286 1287 1288 1289 1290 1291 1292
  forward : maxout(Tensor x, int groups, int axis) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int groups, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : GeneralUnaryGradInferMeta
    param: [x]
  kernel :
    func : maxout_grad

1293
- backward_op : mean_all_grad
Z
zyfncg 已提交
1294 1295 1296 1297 1298 1299 1300 1301 1302
  forward : mean_all(Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : mean_all_grad

1303
- backward_op : mean_double_grad
1304 1305
  forward: mean_grad (Tensor x, Tensor grad_out, IntArray axis={},  bool keepdim=false, bool reduce_all = false) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray axis={},  bool keepdim=false)
Z
zyfncg 已提交
1306
  output : Tensor(grad_out_grad)
1307
  invoke : mean(grad_x_grad, axis, keepdim)
Z
zyfncg 已提交
1308

1309
- backward_op : mean_grad
1310 1311
  forward: mean (Tensor x,  IntArray axis={},  bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray axis={},  bool keepdim=false, bool reduce_all=false)
Z
zyfncg 已提交
1312 1313 1314 1315 1316 1317 1318 1319 1320
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : mean_grad
  backward : mean_double_grad
  no_need_buffer : x

1321
- backward_op : meshgrid_grad
Z
zyfncg 已提交
1322 1323 1324 1325 1326 1327 1328 1329
  forward : meshgrid (Tensor[] inputs) -> Tensor[](outputs)
  args : (Tensor[] inputs, Tensor[] outputs_grad)
  output : Tensor[](inputs_grad){inputs.size()}
  infer_meta :
    func : MeshgridGradInferMeta
  kernel :
    func : meshgrid_grad

1330
- backward_op : min_grad
1331 1332
  forward: min (Tensor x,  IntArray axis={},  bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, IntArray axis={}, bool keepdim=false, bool reduce_all=false)
Z
zyfncg 已提交
1333 1334 1335 1336 1337 1338 1339
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : min_grad

1340
- backward_op : minimum_grad
Z
zyfncg 已提交
1341 1342 1343 1344 1345 1346 1347 1348 1349
  forward : minimum(Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, y]
  kernel :
    func : minimum_grad

1350
- backward_op : mish_grad
Z
zyfncg 已提交
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360
  forward : mish (Tensor x, float threshold) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : mish_grad
  inplace : (out_grad -> x_grad)

1361
- backward_op : mode_grad
Z
zyfncg 已提交
1362 1363 1364 1365 1366 1367 1368 1369 1370
  forward : mode(Tensor x,  int axis,  bool keepdim) -> Tensor(out), Tensor(indices)
  args : (Tensor x, Tensor indices, Tensor out_grad,  int axis,  bool keepdim)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : mode_grad

1371
- backward_op : multi_dot_grad
Z
zyfncg 已提交
1372 1373 1374 1375 1376 1377 1378 1379
  forward : multi_dot (Tensor[] x) -> Tensor(out)
  args : (Tensor[] x, Tensor out_grad)
  output : Tensor[](x_grad) {x.size()}
  infer_meta :
    func : MultiDotGradInferMeta
  kernel :
    func : multi_dot_grad

1380
- backward_op : multiplex_grad
1381 1382 1383
  forward : multiplex (Tensor[] inputs, Tensor index) -> Tensor(out)
  args : (Tensor[] inputs, Tensor index, Tensor out_grad)
  output : Tensor[](inputs_grad){inputs.size()}
Z
zyfncg 已提交
1384 1385
  infer_meta :
    func : MultiplexGradInferMeta
1386
    param : [index, out_grad]
Z
zyfncg 已提交
1387 1388
  kernel :
    func : multiplex_grad
1389
    param : [index, out_grad]
Z
zyfncg 已提交
1390

1391
- backward_op : multiply_double_grad
Z
zyfncg 已提交
1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403
  forward : multiply_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
  args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x, y, grad_out]
  kernel :
    func : multiply_double_grad
  optional : grad_x_grad, grad_y_grad
  backward : multiply_triple_grad
  inplace : (grad_x_grad -> grad_out_grad)

1404
- backward_op : multiply_grad
Z
zyfncg 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
  forward : multiply (Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : multiply_grad
  backward : multiply_double_grad

1415
- backward_op : multiply_triple_grad
Z
zyfncg 已提交
1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
  forward : multiply_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, int aixs = -1) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
  args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
  infer_meta :
    func : GeneralQuinaryGradInferMeta
    param : [x, y, fwd_grad_out, x, y]
  kernel :
    func : multiply_triple_grad
  optional : fwd_grad_grad_x, fwd_grad_grad_y, grad_grad_out_grad

1426
- backward_op : nearest_interp_grad
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
  forward : nearest_interp (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode) -> Tensor(output)
  args : (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, Tensor output_grad, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  optional: out_size, size_tensor, scale_tensor
  kernel :
    func : nearest_interp_grad
    data_type : output_grad

1438
- backward_op : nll_loss_grad
Z
zyfncg 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448
  forward : nll_loss (Tensor input, Tensor label, Tensor weight, int64_t ignore_index, str reduction) -> Tensor(out), Tensor(total_weight)
  args : (Tensor input, Tensor label, Tensor weight, Tensor total_weight, Tensor out_grad, int64_t ignore_index, str reduction)
  output : Tensor(input_grad)
  infer_meta :
    func : NllLossGradInferMeta
  kernel :
    func : nll_loss_grad
    data_type : input
  optional : weight

1449
- backward_op : norm_grad
Z
zyfncg 已提交
1450 1451 1452 1453 1454 1455 1456 1457 1458
  forward : norm (Tensor x, int axis, float epsilon, bool is_test) -> Tensor(out), Tensor(norm)
  args : (Tensor x, Tensor norm, Tensor out_grad, int axis, float epsilon, bool is_test)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : norm_grad

1459
- backward_op : overlap_add_grad
1460 1461 1462 1463 1464 1465 1466 1467 1468
  forward : overlap_add(Tensor x, int hop_length, int axis) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int hop_length, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : OverlapAddGradInferMeta
  kernel :
    func : overlap_add_grad
    data_type : x

1469
- backward_op : p_norm_grad
Z
zyfncg 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478
  forward : p_norm(Tensor x,  float porder,  int axis,  float epsilon,  bool keepdim,  bool asvector=false) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad,  float porder,  int axis,  float epsilon,  bool keepdim,  bool asvector)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : p_norm_grad

1479
- backward_op : pad3d_double_grad
Z
zyfncg 已提交
1480 1481 1482 1483 1484 1485 1486 1487
  forward : pad3d_grad(Tensor x, Tensor grad_out, IntArray paddings, str mode, float pad_value, str data_format) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray paddings, str mode, float pad_value, str data_format)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : Pad3dInferMeta
  kernel :
    func : pad3d

1488
- backward_op : pad3d_grad
Z
zyfncg 已提交
1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499
  forward : pad3d(Tensor x, IntArray paddings, str mode,  float pad_value, str data_format) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray paddings, str mode,  float pad_value, str data_format)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : pad3d_grad
  no_need_buffer : x
  backward : pad3d_double_grad

1500
- backward_op : pad_double_grad
1501 1502
  forward : pad_grad(Tensor x, Tensor grad_out, int[] paddings, Scalar pad_value) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, int[] paddings, Scalar pad_value)
Z
zyfncg 已提交
1503 1504 1505 1506 1507 1508
  output : Tensor(grad_out_grad)
  infer_meta :
    func : PadInferMeta
  kernel :
    func : pad

1509
- backward_op : pad_grad
1510 1511
  forward : pad(Tensor x, int[] paddings, Scalar pad_value) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int[] paddings, Scalar pad_value)
Z
zyfncg 已提交
1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : pad_grad
    param: [out_grad, paddings, pad_value]
  no_need_buffer : x
  backward : pad_double_grad

1522
- backward_op : pixel_shuffle_grad
Z
zyfncg 已提交
1523 1524 1525 1526 1527 1528 1529 1530
  forward : pixel_shuffle (Tensor x, int upscale_factor, str data_format) -> Tensor(out)
  args : (Tensor out_grad, int upscale_factor, str data_format)
  output : Tensor(x_grad)
  infer_meta :
    func : PixelShuffleGradInferMeta
  kernel :
    func : pixel_shuffle_grad

1531
- backward_op : pool2d_double_grad
1532 1533
  forward : pool2d_grad(Tensor x, Tensor out, Tensor grad_out, IntArray kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn)
Z
zyfncg 已提交
1534 1535
  output : Tensor(grad_out_grad)
  infer_meta :
1536
    func : Pool2DInferMeta
1537
    param : [grad_x_grad, kernel_size, strides, paddings, ceil_mode, exclusive, data_format, pooling_type, global_pooling, adaptive, padding_algorithm]
Z
zyfncg 已提交
1538 1539
  kernel :
    func : pool2d_double_grad
1540 1541
    param : [grad_x_grad, kernel_size, strides, paddings, ceil_mode, exclusive, data_format, pooling_type, global_pooling, adaptive, padding_algorithm]
    use_gpudnn : use_gpudnn
Z
zyfncg 已提交
1542

1543
- backward_op : pool2d_grad
1544 1545
  forward : pool2d(Tensor x, IntArray kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, IntArray kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn)
Z
zyfncg 已提交
1546 1547
  output : Tensor(x_grad)
  infer_meta :
1548 1549
    func : UnchangedInferMeta
    param: [x]
Z
zyfncg 已提交
1550 1551
  kernel :
    func : pool2d_grad
1552 1553
    param : [x, out, out_grad, kernel_size, strides, paddings, ceil_mode, exclusive, data_format, pooling_type, global_pooling, adaptive, padding_algorithm]
    use_gpudnn : use_gpudnn
Z
zyfncg 已提交
1554 1555
  backward : pool2d_double_grad

1556
- backward_op : pool3d_grad
1557 1558
  forward : pool3d(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm, bool use_gpudnn)
Z
zyfncg 已提交
1559 1560
  output : Tensor(x_grad)
  infer_meta :
1561 1562
    func : UnchangedInferMeta
    param: [x]
Z
zyfncg 已提交
1563 1564
  kernel :
    func : pool3d_grad
1565 1566
    param : [x, out, out_grad, kernel_size, strides, paddings, ceil_mode, exclusive, data_format, pooling_type, global_pooling, adaptive, padding_algorithm]
    use_gpudnn : use_gpudnn
Z
zyfncg 已提交
1567

1568
- backward_op : pow_grad
1569 1570
  forward : pow(Tensor x, Scalar y) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, Scalar y=-1)
Z
zyfncg 已提交
1571 1572 1573 1574 1575 1576 1577 1578
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : pow_grad
  inplace : (out_grad -> x_grad)

1579
- backward_op : prelu_grad
Z
zyfncg 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588
  forward : prelu(Tensor x, Tensor alpha, str data_format, str mode) -> Tensor(out)
  args : (Tensor x, Tensor alpha, Tensor out_grad, str data_format, str mode)
  output : Tensor(x_grad), Tensor(alpha_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param: [x, alpha]
  kernel :
    func : prelu_grad

1589
- backward_op : psroi_pool_grad
Z
zyfncg 已提交
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601
  forward : psroi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, int output_channels, float spatial_scale) -> Tensor(out)
  args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, int output_channels, float spatial_scale)
  output : Tensor(x_grad)
  infer_meta :
    func : GeneralUnaryGradInferMeta
    param : [x]
  kernel :
    func : psroi_pool_grad
    data_type : x
  optional : boxes_num

# output is optional
1602
- backward_op : put_along_axis_grad
1603 1604
  forward : put_along_axis (Tensor arr, Tensor indices, Tensor value, int axis, str reduce) -> Tensor(out)
  args : (Tensor arr, Tensor indices, Tensor out_grad, int axis, str reduce)
1605
  output : Tensor(arr_grad), Tensor(value_grad)
Z
zyfncg 已提交
1606 1607
  infer_meta :
    func : GeneralBinaryGradInferMeta
1608
    param : [arr, indices]
Z
zyfncg 已提交
1609 1610 1611
  kernel :
    func : put_along_axis_grad

1612
- backward_op : qr_grad
Y
Yulong Ao 已提交
1613 1614 1615 1616 1617 1618 1619 1620 1621
  forward : qr (Tensor x, str mode) -> Tensor(q), Tensor(r)
  args : (Tensor x, Tensor q, Tensor r, Tensor q_grad, Tensor r_grad, str mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : qr_grad

1622
- backward_op : real_grad
Z
zyfncg 已提交
1623 1624 1625 1626 1627
  forward : real (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  invoke : real_grad_impl(out_grad, x_grad)

1628
- backward_op : reciprocal_grad
Z
zyfncg 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637 1638
  forward : reciprocal (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : reciprocal_grad
  inplace : (out_grad -> x_grad)

1639
- backward_op : reduce_prod_grad
1640 1641
  forward : reduce_prod (Tensor x, IntArray dims, bool keep_dim, bool reduce_all) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad, IntArray dims,  bool keep_dim, bool reduce_all)
Z
zyfncg 已提交
1642 1643 1644 1645 1646 1647 1648
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : prod_grad

1649
- backward_op : relu6_grad
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659
  forward : relu6 (Tensor x, float threshold) -> Tensor(out)
  args : (Tensor out, Tensor out_grad, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : relu6_grad
  inplace : (out_grad -> x_grad)

1660
- backward_op : relu_double_grad
Z
zyfncg 已提交
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670
  forward : relu_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor out, Tensor grad_x_grad)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : relu_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

1671
- backward_op : relu_grad
Z
zyfncg 已提交
1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682
  forward : relu (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : relu_grad
  backward: relu_double_grad
  inplace : (out_grad -> x_grad)

1683
- backward_op : renorm_grad
S
seemingwang 已提交
1684 1685 1686 1687 1688 1689 1690 1691 1692
  forward : renorm (Tensor x, float p, int axis, float max_norm) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float p, int axis, float max_norm)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : renorm_grad

1693
- backward_op : repeat_interleave_grad
1694 1695
  forward : repeat_interleave(Tensor x, int repeats, int axis) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int repeats, int axis)
S
seemingwang 已提交
1696 1697 1698 1699 1700 1701 1702
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : repeat_interleave_grad

1703
- backward_op : repeat_interleave_with_tensor_index_grad
1704 1705
  forward : repeat_interleave_with_tensor_index(Tensor x, Tensor repeats, int axis) -> Tensor(out)
  args : (Tensor x, Tensor repeats, Tensor out_grad, int axis)
S
seemingwang 已提交
1706 1707 1708 1709 1710 1711 1712 1713
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : repeat_interleave_with_tensor_index_grad
    data_type : x

1714
- backward_op : reshape_double_grad
Z
zyfncg 已提交
1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
  forward : reshape_grad (Tensor xshape, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor grad_out, Tensor grad_x_grad)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [grad_out]
  kernel :
    func : reshape_double_grad
  no_need_buffer : grad_out
  inplace : (grad_x_grad -> grad_out_grad)

1726
- backward_op : reshape_grad
Z
zyfncg 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
  forward : reshape (Tensor x, IntArray shape) -> Tensor(out), Tensor(xshape)
  args : (Tensor xshape, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : KernelWithXShapeInferMeta
    param : [xshape]
  kernel :
    func : reshape_grad
    param : [out_grad]
    data_type: out_grad
    backend: out_grad
    layout: out_grad
  backward : reshape_double_grad
  inplace : (out_grad -> x_grad)

1742
- backward_op : reverse_array_grad
1743 1744
  forward : reverse_array (Tensor[] x, IntArray axis) -> Tensor[](out)
  args : (Tensor[] out_grad, IntArray axis)
W
wanghuancoder 已提交
1745 1746 1747 1748 1749 1750
  output : Tensor[](x_grad){out_grad.size()}
  infer_meta :
    func : ReverseArrayInferMeta
  kernel :
    func : reverse

1751
- backward_op : reverse_grad
1752 1753
  forward : reverse (Tensor x, IntArray axis) -> Tensor(out)
  args : (Tensor out_grad, IntArray axis)
W
wanghuancoder 已提交
1754 1755 1756
  output : Tensor(x_grad)
  invoke : reverse(out_grad, axis)

Y
YuanRisheng 已提交
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768
- backward_op : rnn_grad
  forward : rnn (Tensor x, Tensor[] pre_state, Tensor[] weight_list, Tensor sequence_length, Tensor dropout_state_in, float dropout_prob, bool is_bidirec, int input_size, int hidden_size, int num_layers, str mode, int seed, bool is_test) -> Tensor(out), Tensor(dropout_state_out), Tensor[](state), Tensor(reserve)
  args : (Tensor x, Tensor[] pre_state, Tensor[] weight_list, Tensor sequence_length, Tensor out, Tensor dropout_state_out, Tensor reserve, Tensor out_grad, Tensor[] state_grad, float dropout_prob, bool is_bidirec, int input_size, int hidden_size, int num_layers, str mode, int seed, bool is_test)
  output : Tensor(x_grad), Tensor[](pre_state_grad){pre_state.size()}, Tensor[](weight_list_grad){weight_list.size()}
  infer_meta :
    func : RnnGradInferMeta
    param : [x, pre_state, weight_list]
  kernel :
    func : rnn_grad
    data_type: out_grad
  optional : sequence_length

1769
- backward_op : roi_align_grad
Z
zyfncg 已提交
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781
  forward : roi_align (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned) -> Tensor(out)
  args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : roi_align_grad
    data_type : boxes
  no_need_buffer : x
  optional : boxes_num

1782
- backward_op : roi_pool_grad
Z
zyfncg 已提交
1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793
  forward : roi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale) -> Tensor(out), Tensor(arg_max)
  args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor arg_max, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : roi_pool_grad
    data_type : x
  optional : boxes_num

1794
- backward_op : roll_grad
Z
zyfncg 已提交
1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
  forward : roll(Tensor x, IntArray shifts, int64_t[] axis) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray shifts, int64_t[] axis)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : roll_grad
    data_type : x
  no_need_buffer : x

1806
- backward_op : round_grad
Z
zyfncg 已提交
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816
  forward : round(Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [out_grad]
  kernel :
    func : round_grad
  inplace : (out_grad -> x_grad)

1817
- backward_op : rsqrt_double_grad
Z
zyfncg 已提交
1818 1819 1820 1821 1822 1823 1824 1825 1826 1827
  forward : rsqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
  output : Tensor(out_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [out, out]
  kernel :
    func : rsqrt_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

1828
- backward_op : rsqrt_grad
Z
zyfncg 已提交
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839
  forward : rsqrt (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : rsqrt_grad
  backward : rsqrt_double_grad
  inplace : (out_grad -> x_grad)

1840
- backward_op : scale_grad
Z
zyfncg 已提交
1841
  forward : scale (Tensor x, Scalar scale, float bias, bool bias_after_scale) -> Tensor(out)
1842
  args : (Tensor out_grad, Scalar scale=1.0, bool bias_after_scale=true)
Z
zyfncg 已提交
1843 1844 1845
  output : Tensor(x_grad)
  invoke : scale(out_grad, scale, 0.0, bias_after_scale)

1846
- backward_op : scatter_grad
Z
zyfncg 已提交
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856
  forward : scatter (Tensor x, Tensor index, Tensor updates, bool overwrite) -> Tensor(out)
  args : (Tensor index, Tensor updates, Tensor out_grad, bool overwrite)
  output : Tensor(x_grad), Tensor(updates_grad)
  infer_meta :
    func : ScatterGradInferMeta
    param : [index, updates, out_grad, overwrite]
  kernel :
    func : scatter_grad
  no_need_buffer : updates

1857
- backward_op : scatter_nd_add_grad
Z
zyfncg 已提交
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867
  forward : scatter_nd_add (Tensor x, Tensor index, Tensor updates) -> Tensor(out)
  args : (Tensor index, Tensor updates, Tensor out_grad)
  output : Tensor(x_grad), Tensor(updates_grad)
  infer_meta :
    func : ScatterNdAddGradInferMeta
    param : [index, updates, out_grad]
  kernel :
    func : scatter_nd_add_grad
  no_need_buffer : updates

1868
- backward_op : segment_pool_grad
Z
zyfncg 已提交
1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
  forward : segment_pool (Tensor x, Tensor segment_ids, str pooltype) -> Tensor(out), Tensor(summed_ids)
  args : (Tensor x, Tensor segment_ids, Tensor out, Tensor summed_ids, Tensor out_grad, str pooltype)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : segment_pool_grad
    data_type : x
  optional : summed_ids

1880
- backward_op : selu_grad
Z
zyfncg 已提交
1881 1882 1883 1884 1885 1886 1887 1888 1889
  forward : selu (Tensor x, float scale, float alpha) -> Tensor(out)
  args : (Tensor out, Tensor out_grad, float scale, float alpha)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : selu_grad

1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913
- backward_op : send_u_recv_grad
  forward : send_u_recv (Tensor x, Tensor src_index, Tensor dst_index, str reduce_op = "SUM", IntArray out_size = {0}) -> Tensor(out), Tensor(dst_count)
  args : (Tensor x, Tensor src_index, Tensor dst_index, Tensor out, Tensor dst_count, Tensor out_grad, str reduce_op = "SUM")
  output : Tensor(x_grad)
  infer_meta :
    func : GeneralUnaryGradInferMeta
    param : [x]
  kernel :
    func : send_u_recv_grad
    data_type : out_grad
  optional: out, dst_count

- backward_op : send_ue_recv_grad
  forward : send_ue_recv (Tensor x, Tensor y, Tensor src_index, Tensor dst_index, str message_op, str reduce_op, IntArray out_size) -> Tensor(out), Tensor(dst_count)
  args : (Tensor x, Tensor y, Tensor src_index, Tensor dst_index, Tensor out, Tensor dst_count, Tensor out_grad, str message_op, str reduce_op)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : send_ue_recv_grad
    data_type : out_grad
  optional: out, dst_count

1914
- backward_op : sigmoid_cross_entropy_with_logits_grad
Z
zyfncg 已提交
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
  forward : sigmoid_cross_entropy_with_logits (Tensor x, Tensor label, bool normalize, int ignore_index) -> Tensor(out)
  args : (Tensor x, Tensor label, Tensor out_grad, bool normalize, int ignore_index)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : sigmoid_cross_entropy_with_logits_grad
  inplace : (out_grad -> x_grad)

1925
- backward_op : sigmoid_double_grad
Z
zyfncg 已提交
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936
  forward : sigmoid_grad (Tensor out, Tensor fwd_grad_out) -> Tensor(grad_x)
  args : (Tensor out, Tensor fwd_grad_out, Tensor grad_x_grad)
  output : Tensor(out_grad), Tensor(fwd_grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [out, fwd_grad_out]
  kernel :
    func : sigmoid_double_grad
  backward : sigmoid_triple_grad
  inplace : (grad_x_grad -> fwd_grad_out_grad)

1937
- backward_op : sigmoid_grad
Z
zyfncg 已提交
1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948
  forward : sigmoid (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : sigmoid_grad
  backward : sigmoid_double_grad
  inplace : (out_grad -> x_grad)

1949
- backward_op : sigmoid_triple_grad
Z
zyfncg 已提交
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960
  forward : sigmoid_double_grad (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x) -> Tensor(grad_out), Tensor(grad_grad_out)
  args : (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x, Tensor grad_out_grad, Tensor grad_grad_out_grad)
  output : Tensor(out_grad), Tensor(fwd_grad_out_grad), Tensor(grad_grad_x_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [out, fwd_grad_out, grad_grad_x]
  kernel :
    func : sigmoid_triple_grad
  optional : grad_grad_out_grad
  inplace : (grad_grad_x -> fwd_grad_out_grad)

1961 1962 1963 1964 1965 1966
- backward_op : sign_grad
  forward : sign (Tensor x) -> Tensor(out)
  args : (Tensor out_grad)
  output : Tensor(x_grad)
  invoke : scale(out_grad, 0.0, 0.0, true)

1967
- backward_op : silu_grad
Z
zyfncg 已提交
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
  forward : silu (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : silu_grad
  inplace : (out_grad -> x_grad)

1978
- backward_op : slice_double_grad
1979 1980 1981
  forward : slice_grad (Tensor input, Tensor grad_out, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis) -> Tensor(grad_input)
  args : (Tensor grad_input_grad, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
  output : Tensor(grad_out_grad)
1982
  invoke : slice(grad_input_grad, axes, starts, ends, infer_flags, decrease_axis)
1983

1984
- backward_op : slice_grad
Z
zyfncg 已提交
1985 1986 1987 1988 1989 1990 1991 1992
  forward : slice (Tensor input, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis) -> Tensor(out)
  args : (Tensor input, Tensor out_grad, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
  output : Tensor(input_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [input]
  kernel :
    func : slice_grad
1993
  backward : slice_double_grad
Z
zyfncg 已提交
1994 1995
  no_need_buffer : input

1996
- backward_op : slogdet_grad
1997 1998 1999 2000 2001 2002 2003 2004 2005
  forward : slogdet (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : slogdeterminant_grad

2006
- backward_op : softmax_grad
Z
zyfncg 已提交
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
  forward : softmax (Tensor x, int axis) -> Tensor(out)
  args : (Tensor out, Tensor out_grad, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : softmax_grad
    use_gpudnn : true

2017
- backward_op : softplus_grad
W
Wang Bojun 已提交
2018 2019 2020 2021 2022 2023 2024 2025 2026 2027
  forward : softplus (Tensor x, float beta, float threshold) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float beta, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : softplus_grad
  inplace : (out_grad -> x_grad)

2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038
- backward_op : softshrink_grad
  forward : softshrink (Tensor x, float threshold) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : soft_shrink_grad
  inplace : (out_grad -> x_grad)

2039
- backward_op : softsign_grad
2040 2041 2042 2043 2044 2045 2046 2047 2048 2049
  forward : softsign (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : softsign_grad
  inplace : (out_grad -> x_grad)

2050
- backward_op : spectral_norm_grad
2051 2052 2053 2054 2055 2056 2057 2058 2059
  forward : spectral_norm (Tensor weight, Tensor u, Tensor v, int dim, int power_iters, float eps) -> Tensor(out)
  args : (Tensor weight, Tensor u, Tensor v, Tensor out_grad, int dim, int power_iters, float eps)
  output : Tensor(weight_grad)
  infer_meta :
    func : SpectralNormGradInferMeta
  kernel :
    func : spectral_norm_grad
    data_type : out_grad

2060
- backward_op : split_grad
Z
zyfncg 已提交
2061 2062 2063 2064
  forward : split (Tensor x, IntArray num_or_sections, Scalar axis) -> Tensor[](out)
  args : (Tensor[] out_grad, Scalar axis = -1)
  output : Tensor(x_grad)
  invoke : concat( out_grad, axis)
C
Charles-hit 已提交
2065

2066
- backward_op : split_with_num_grad
C
Charles-hit 已提交
2067 2068 2069 2070
  forward : split_with_num (Tensor x, int num, Scalar axis) -> Tensor[](out)
  args : (Tensor[] out_grad, Scalar axis = -1)
  output : Tensor(x_grad)
  invoke : concat( out_grad, axis)
Z
zyfncg 已提交
2071 2072
# TODO(zhangyunfei) The config of double grad and triple grad will be supported in the future.

2073
- backward_op : sqrt_double_grad
Z
zyfncg 已提交
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083
  forward : sqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
  output : Tensor(out_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [out, out]
  kernel :
    func : sqrt_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

2084
- backward_op : sqrt_grad
Z
zyfncg 已提交
2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095
  forward : sqrt (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : sqrt_grad
  backward : sqrt_double_grad
  inplace : (out_grad -> x_grad)

2096
- backward_op : square_double_grad
Z
zyfncg 已提交
2097 2098 2099 2100 2101 2102 2103 2104 2105 2106
  forward : square_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
  output : Tensor(x_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, x]
  kernel :
    func : square_double_grad
  inplace : (grad_x_grad -> grad_out_grad)

2107
- backward_op : square_grad
Z
zyfncg 已提交
2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118
  forward : square (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : square_grad
  backward : square_double_grad
  inplace : (out_grad -> x_grad)

2119
- backward_op : squared_l2_norm_grad
2120 2121 2122 2123 2124 2125 2126 2127 2128
  forward : squared_l2_norm(Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  kernel :
    func : squared_l2_norm_grad

2129
- backward_op : squeeze_double_grad
2130 2131
  forward : squeeze_grad(Tensor xshape, Tensor grad_out, IntArray axis) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray axis)
Z
zyfncg 已提交
2132
  output : Tensor(grad_out_grad)
2133
  invoke: squeeze(grad_x_grad, axis)
Z
zyfncg 已提交
2134

2135
- backward_op : squeeze_grad
2136 2137
  forward : squeeze(Tensor x, IntArray axis) -> Tensor(out), Tensor(xshape)
  args : (Tensor xshape, Tensor out_grad, IntArray axis)
Z
zyfncg 已提交
2138 2139 2140 2141 2142 2143 2144 2145 2146
  output : Tensor(x_grad)
  infer_meta :
    func : KernelWithXShapeInferMeta
    param: [xshape]
  kernel :
    func : squeeze_grad
  inplace : (out_grad -> x_grad)
  backward: squeeze_double_grad

2147
- backward_op : stack_grad
Z
zyfncg 已提交
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158
  forward : stack (Tensor[] x, int axis) -> Tensor(out)
  args : (Tensor[] x, Tensor out_grad, int axis)
  output : Tensor[](x_grad){x.size()}
  infer_meta :
    func : StackGradInferMeta
    param: [out_grad, axis]
  kernel :
    func : stack_grad
    param : [out_grad, axis]
  no_need_buffer : x

2159
- backward_op : strided_slice_grad
Z
zyfncg 已提交
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169
  forward : strided_slice (Tensor x, int[] axes, IntArray starts, IntArray ends, IntArray strides) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int[] axes, IntArray starts, IntArray ends, IntArray strides)
  output : Tensor(x_grad)
  infer_meta :
    func : GeneralUnaryGradInferMeta
    param : [x]
  kernel :
    func : strided_slice_grad
  no_need_buffer : x

2170
- backward_op : subtract_double_grad
Z
zyfncg 已提交
2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182
  forward : subtract_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
  args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
  output : Tensor(grad_out_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [grad_out]
  kernel :
    func : subtract_double_grad
  optional : grad_x_grad, grad_y_grad
  no_need_buffer : y, grad_out
  inplace : (grad_x_grad -> grad_out_grad)

2183
- backward_op : subtract_grad
Z
zyfncg 已提交
2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195
  forward : subtract (Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : subtract_grad
  no_need_buffer : x, y
  backward : subtract_double_grad
  inplace : (out_grad -> x_grad)

2196
- backward_op : sum_double_grad
2197 2198
  forward : sum_grad (Tensor x, Tensor grad_out, IntArray axis, bool keepdim, bool reduce_all=false) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray axis={}, bool keepdim=false)
Z
zyfncg 已提交
2199
  output : Tensor(grad_out_grad)
2200
  invoke : sum(grad_x_grad, axis, grad_x_grad.dtype(), keepdim)
Z
zyfncg 已提交
2201

2202
- backward_op : sum_grad
2203 2204
  forward : sum (Tensor x, IntArray axis={}, DataType dtype=DataType::UNDEFINED, bool keepdim=false) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray axis, bool keepdim, bool reduce_all=false)
Z
zyfncg 已提交
2205 2206 2207 2208 2209 2210 2211 2212 2213
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : sum_grad
  no_need_buffer : x
  backward : sum_double_grad

2214
- backward_op : svd_grad
2215 2216
  forward : svd (Tensor x, bool full_matrices) -> Tensor(u), Tensor(s), Tensor(vh)
  args : (Tensor x, Tensor u, Tensor vh, Tensor s, Tensor u_grad, Tensor vh_grad, Tensor s_grad, bool full_matrices)
2217 2218 2219 2220 2221 2222 2223 2224
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : svd_grad
  optional: u_grad, vh_grad, s_grad

2225
- backward_op : swish_grad
Z
zyfncg 已提交
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235
  forward : swish (Tensor x, float beta=1.0) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float bete=1.0)
  output : Tensor(x_grad)
  infer_meta :
    func : GeneralUnaryGradInferMeta
    param : [x]
  kernel :
    func : swish_grad
  inplace : (out_grad -> x_grad)

2236
- backward_op : sync_batch_norm_grad
2237
  forward : sync_batch_norm_ (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
2238
  args : (Tensor x, Tensor scale, Tensor bias, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor out_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
2239 2240 2241 2242 2243 2244 2245
  output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [x, scale, bias]
  kernel :
    func : sync_batch_norm_grad
    data_type : out_grad
2246
  optional : reserve_space
2247

2248
- backward_op : take_along_axis_grad
2249 2250 2251
  forward : take_along_axis (Tensor arr, Tensor indices, int axis) -> Tensor(out)
  args : (Tensor arr, Tensor indices, Tensor out_grad, int axis)
  output : Tensor(arr_grad)
Z
zyfncg 已提交
2252 2253
  infer_meta :
    func : UnchangedInferMeta
2254
    param : [arr]
Z
zyfncg 已提交
2255 2256 2257
  kernel :
    func : take_along_axis_grad

2258
- backward_op : tanh_double_grad
Z
zyfncg 已提交
2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269
  forward : tanh_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
  args : (Tensor out, Tensor grad_out, Tensor grad_x_grad)
  output : Tensor(out_grad), Tensor(grad_out_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [out, out]
  kernel :
    func : tanh_double_grad
  backward : tanh_triple_grad
  inplace : (grad_x_grad -> grad_out_grad)

2270
- backward_op : tanh_grad
Z
zyfncg 已提交
2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
  forward : tanh (Tensor x) -> Tensor(out)
  args : (Tensor out, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out]
  kernel :
    func : tanh_grad
  backward : tanh_double_grad
  inplace : (out_grad -> x_grad)

2282
- backward_op : tanh_shrink_grad
Z
zyfncg 已提交
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292
  forward : tanh_shrink (Tensor x) -> Tensor(out)
  args : (Tensor x, Tensor out_grad)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : tanh_shrink_grad
  inplace : (out_grad -> x_grad)

2293
- backward_op : tanh_triple_grad
Z
zyfncg 已提交
2294 2295 2296 2297 2298 2299 2300 2301 2302 2303
  forward : tanh_double_grad (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward) -> Tensor(grad_out_new), Tensor(grad_out_grad)
  args : (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward, Tensor grad_out_new_grad, Tensor grad_out_grad_grad)
  output : Tensor(out_grad), Tensor(grad_out_forward_grad), Tensor(grad_x_grad_forward_grad)
  infer_meta :
    func : GeneralTernaryGradInferMeta
    param : [out, out, grad_x_grad_forward]
  kernel :
    func : tanh_triple_grad
  inplace : (grad_x_grad_forward -> grad_out_forward_grad)

2304
- backward_op : temporal_shift_grad
C
ccrrong 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313
  forward : temporal_shift(Tensor x, int seg_num, float shift_ratio, str data_format_str) -> Tensor(out)
  args : (Tensor out_grad, int seg_num, float shift_ratio, str data_format_str)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : temporal_shift_grad

2314
- backward_op : thresholded_relu_grad
Z
zyfncg 已提交
2315 2316 2317 2318 2319 2320 2321 2322 2323 2324
  forward : thresholded_relu (Tensor x, float threshold) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, float threshold)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : thresholded_relu_grad
  inplace : (out_grad -> x_grad)

2325
- backward_op : tile_double_grad
Z
zyfncg 已提交
2326 2327 2328
  forward : tile_grad (Tensor x, Tensor grad_out, IntArray repeat_times) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray repeat_times)
  output : Tensor(grad_out_grad)
2329
  invoke : tile(grad_x_grad, repeat_times)
Z
zyfncg 已提交
2330

2331
- backward_op : tile_grad
Z
zyfncg 已提交
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342
  forward : tile (Tensor x, IntArray repeat_times) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, IntArray repeat_times)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : tile_grad
  no_need_buffer : x
  backward : tile_double_grad

2343 2344
- backward_op : topk_grad
  forward : topk (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true) -> Tensor(out), Tensor(indices)
Z
zyfncg 已提交
2345 2346 2347 2348 2349 2350 2351 2352
  args : (Tensor x, Tensor indices, Tensor out_grad, Scalar k = -1, int axis = -1, bool largest = true, bool sorted = true)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : top_k_grad

2353
- backward_op : transpose_double_grad
2354 2355
  forward : transpose_grad (Tensor grad_out, int[] perm) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, int[] perm)
Z
zyfncg 已提交
2356
  output : Tensor(grad_out_grad)
2357
  invoke : transpose(grad_x_grad, perm)
Z
zyfncg 已提交
2358

2359
- backward_op : transpose_grad
2360 2361
  forward : transpose (Tensor x, int[] perm) -> Tensor(out)
  args : (Tensor out_grad, int[] perm)
Z
zyfncg 已提交
2362 2363 2364
  output : Tensor(x_grad)
  infer_meta :
    func : TransposeGradInferMeta
2365
    param : [out_grad, perm]
Z
zyfncg 已提交
2366 2367 2368 2369
  kernel :
    func : transpose_grad
  backward : transpose_double_grad

2370
- backward_op : triangular_solve_grad
Z
zyfncg 已提交
2371 2372 2373 2374 2375 2376 2377 2378 2379
  forward : triangular_solve (Tensor x, Tensor y, bool upper, bool tranpose, bool unitriangular) -> Tensor(out)
  args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, bool upper, bool tranpose, bool unitriangular)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : triangular_solve_grad

2380
- backward_op : tril_triu_grad
Z
zyfncg 已提交
2381 2382 2383 2384 2385 2386 2387 2388 2389
  forward : tril_triu(Tensor x,  int diagonal,  bool lower) -> Tensor(out)
  args : (Tensor out_grad,  int diagonal,  bool lower)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [out_grad]
  kernel :
    func : tril_triu_grad

2390
- backward_op : trilinear_interp_grad
2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401
  forward : trilinear_interp (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode) -> Tensor(output)
  args : (Tensor x, Tensor out_size, Tensor[] size_tensor, Tensor scale_tensor, Tensor output_grad, str data_layout, int out_d, int out_h, int out_w, float[] scale, str interp_method, bool align_corners, int align_mode)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param: [x]
  optional: out_size, size_tensor, scale_tensor
  kernel :
    func : trilinear_interp_grad
    data_type : output_grad

2402
- backward_op : unbind_grad
Z
zyfncg 已提交
2403 2404 2405 2406 2407
  forward : unbind (Tensor input, int axis) -> Tensor[](out)
  args : (Tensor[] out_grad, int axis)
  output : Tensor(input_grad)
  invoke : stack(out_grad, axis)

2408
- backward_op : unfold_grad
Z
zyfncg 已提交
2409 2410 2411 2412 2413 2414 2415 2416 2417 2418
  forward : unfold (Tensor x, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations) -> Tensor(out)
  args : (Tensor x, Tensor out_grad, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
  output : Tensor(x_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [x]
  kernel :
    func : unfold_grad
  no_need_buffer : x

2419
- backward_op : uniform_random_inplace_grad
2420 2421 2422 2423 2424 2425 2426 2427 2428
  forward : uniform_random_inplace(Tensor x, float min, float max, int seed, int diag_num, int diag_step, float diag_val) -> Tensor(out)
  args : (Tensor out_grad, float min, float max, int seed, int diag_num, int diag_step, float diag_val)
  output : Tensor(x_grad)
  infer_meta :
    func : UniformRandomInplaceGradInferMeta
  kernel :
    func : uniform_random_inplace_grad
  inplace : (out_grad -> x_grad)

2429
- backward_op : unsqueeze_double_grad
Z
zyfncg 已提交
2430 2431 2432 2433 2434
  forward : unsqueeze_grad(Tensor xshape, Tensor grad_out, IntArray axes) -> Tensor(grad_x)
  args : (Tensor grad_x_grad, IntArray axes)
  output : Tensor(grad_out_grad)
  invoke : unsqueeze(grad_x_grad, axes)

2435
- backward_op : unsqueeze_grad
Z
zyfncg 已提交
2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447
  forward : unsqueeze(Tensor x, IntArray axes) -> Tensor(out), Tensor(xshape)
  args : (Tensor xshape, Tensor out_grad, IntArray axes)
  output : Tensor(x_grad)
  infer_meta :
    func : KernelWithXShapeInferMeta
    param: [xshape]
  kernel :
    func : unsqueeze_grad
    param: [xshape, out_grad]
  inplace : (out_grad -> x_grad)
  backward : unsqueeze_double_grad

2448
- backward_op : unstack_grad
2449 2450 2451 2452 2453 2454 2455 2456 2457
  forward : unstack (Tensor x, int axis, int num) -> Tensor[](out)
  args : (Tensor[] out_grad, int axis)
  output : Tensor(x_grad)
  infer_meta :
    func : UnStackGradInferMeta
    param : [out_grad, axis]
  kernel :
    func : unstack_grad

2458
- backward_op : warpctc_grad
2459
  forward : warpctc (Tensor logits, Tensor label, Tensor logits_length, Tensor labels_length, int blank, bool norm_by_times) -> Tensor(loss), Tensor(warpctcgrad)
Z
Zhong Hui 已提交
2460 2461 2462 2463 2464 2465 2466 2467 2468 2469
  args : (Tensor logits, Tensor logits_length, Tensor warpctcgrad, Tensor loss_grad, int blank, bool norm_by_times)
  output : Tensor(logits_grad)
  infer_meta :
    func : UnchangedInferMeta
    param : [logits]
  kernel :
    func : warpctc_grad
  optional : logits_length
  no_need_buffer : logits

2470
- backward_op : where_grad
Z
zyfncg 已提交
2471 2472 2473 2474 2475 2476 2477 2478 2479
  forward : where (Tensor condition, Tensor x, Tensor y) -> Tensor(out)
  args : (Tensor condition, Tensor x, Tensor y, Tensor out_grad)
  output : Tensor(x_grad), Tensor(y_grad)
  infer_meta :
    func : GeneralBinaryGradInferMeta
    param : [x, y]
  kernel :
    func : where_grad
  no_need_buffer : x, y
2480

2481
- backward_op : yolov3_loss_grad
2482 2483 2484 2485 2486 2487 2488 2489
  forward : yolov3_loss(Tensor x, Tensor gt_box, Tensor gt_label, Tensor gt_score, int[] anchors, int[] anchor_mask, int class_num, float ignore_thresh, int downsample_ratio, bool use_label_smooth=true, float scale_x_y=1.0) -> Tensor(loss), Tensor(objectness_mask), Tensor(gt_match_mask)
  args : (Tensor x, Tensor gt_box, Tensor gt_label, Tensor gt_score, Tensor objectness_mask, Tensor gt_match_mask, Tensor loss_grad, int[] anchors, int[] anchor_mask, int class_num, float ignore_thresh, int downsample_ratio, bool use_label_smooth=true, float scale_x_y=1.0)
  output : Tensor(x_grad), Tensor(gt_box_grad), Tensor(gt_label_grad), Tensor(gt_score_grad)
  infer_meta :
    func : Yolov3LossGradInferMeta
  kernel :
    func : yolov3_loss_grad
  optional : gt_score
X
xiaoting 已提交
2490

2491
- backward_op: fold_grad
X
xiaoting 已提交
2492 2493 2494 2495 2496 2497 2498 2499 2500 2501
  forward: fold (Tensor x, int[] output_sizes, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations) -> Tensor(out)
  args: (Tensor x, Tensor out_grad, int[] output_sizes, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
  output: Tensor(x_grad)
  infer_meta:
    func: UnchangedInferMeta
    param : [x]
  kernel:
    func: fold_grad
  no_need_buffer : x

2502
- backward_op: unpool3d_grad
X
xiaoting 已提交
2503 2504 2505 2506 2507 2508 2509 2510 2511 2512
  forward: unpool3d (Tensor x, Tensor indices, int[] ksize, int[] strides, int[] padding, int[] output_size, str data_format) -> Tensor(out)
  args: (Tensor x, Tensor indices, Tensor out, Tensor out_grad, int[] ksize, int[] strides, int[] padding, int[] output_size, str data_format)
  output: Tensor(x_grad)
  infer_meta:
    func: UnchangedInferMeta
    param : [x]
  kernel:
    func: unpool3d_grad
    data_type: x

2513
- backward_op: unpool_grad
2514 2515
  forward: unpool (Tensor x, Tensor indices, int[] ksize, int[] strides, int[] padding,  IntArray output_size, str data_format) -> Tensor(out)
  args: (Tensor x, Tensor indices, Tensor out, Tensor out_grad, int[] ksize, int[] strides, int[] padding, IntArray output_size, str data_format)
X
xiaoting 已提交
2516 2517 2518 2519 2520 2521 2522
  output: Tensor(x_grad)
  infer_meta:
    func: UnchangedInferMeta
    param : [x]
  kernel:
    func: unpool_grad
    data_type: x