A. Unique TensorFlower 0e4a3676d9 Clean up duplicated code in matmul_op/batch_matmul_op and associated tests. Having a separate redundant implementation for non-batch matmul is a maintenance burden. ... We add a registration of the batch kernel under the "MatMul" symbol (with modified shape validation) for legacy graphs. PiperOrigin-RevId: 339715052 Change-Id: Iad428cdf62a656de56f62d996db44be682af5d13		2020-10-29 11:58:12 -07:00
..
banded_triangular_solve_op_test.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
banded_triangular_solve_op.cc	Prevent integer truncation from 64 to 32 bits.	2020-09-18 17:56:30 -07:00
BUILD	Remove depencencies on unnecessary aliases in tensorflow/core	2020-10-13 10:28:00 -07:00
cholesky_grad.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
cholesky_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
determinant_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
determinant_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
determinant_op.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eig_op_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eig_op_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eig_op_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eig_op_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eig_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_bfloat16.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_half.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_int32.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl_int64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
einsum_op_impl.h	Clean up duplicated code in matmul_op/batch_matmul_op and associated tests. Having a separate redundant implementation for non-batch matmul is a maintenance burden.	2020-10-29 11:58:12 -07:00
einsum_op.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eye_functor_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
eye_functor.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
linalg_ops_common.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
linalg_ops_common.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
lu_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
lu_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_band_part_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_band_part_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_band_part_op.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_diag_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_diag_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_diag_op.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_exponential_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_inverse_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_logarithm_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_set_diag_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_set_diag_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_set_diag_op.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_ls_op_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_ls_op_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_ls_op_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_ls_op_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_ls_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_solve_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_square_root_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_triangular_solve_op_complex.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_triangular_solve_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_triangular_solve_op_real.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
matrix_triangular_solve_op_test.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
qr_op_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
qr_op_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
qr_op_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
qr_op_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
qr_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_gpu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
self_adjoint_eig_v2_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
svd_op_complex64.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
svd_op_complex128.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
svd_op_double.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
svd_op_float.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
svd_op_gpu.cu.cc	Verify dimensions before we assume we have at least 2 (i.e. before we end up calling dim_size with a negative dim index). This fixes tensorflow/python/kernel_tests:svd_op_test_gpu when running with asan.	2020-10-19 19:12:26 -07:00
svd_op_impl.h	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
tridiagonal_matmul_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
tridiagonal_matmul_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
tridiagonal_solve_op_gpu.cu.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00
tridiagonal_solve_op.cc	Break up core/kernels/BUILD (part 1 of N):	2020-08-04 17:24:28 -07:00