CUDA: masked_scale_cuda
- func: _reshape_from_tensor(Tensor self, Tensor shape) -> Tensor
+ matches_jit_signature: True
- func: _shape_as_tensor(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: dropout(Tensor input, double p, bool train) -> Tensor
- func: feature_alpha_dropout_(Tensor self, double p, bool train) -> Tensor
- func: abs(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: abs_(Tensor self) -> Tensor
CUDA: _abs_out_cuda
- func: acos(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: acos_(Tensor self) -> Tensor
- func: adaptive_max_pool1d(Tensor self, IntList[1] output_size) -> (Tensor, Tensor)
- func: add(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: add_(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
# For C++ only, until we have conversion from C++ numbers to Tensor
- func: add(Tensor self, Scalar other, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: add_(Tensor self, Scalar other, Scalar alpha=1) -> Tensor
variants: method
- func: addmv(Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: addmv_(Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1) -> Tensor
- func: addmv_out(Tensor result, Tensor self, Tensor mat, Tensor vec, *, Scalar beta=1, Scalar alpha=1) -> Tensor
- func: addr(Tensor self, Tensor vec1, Tensor vec2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: addr_(Tensor self, Tensor vec1, Tensor vec2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
variants: function, method
- func: argmax(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: _argmax(Tensor self, int64_t dim, bool keepdim=false) -> Tensor
variants: function, method
- func: argmin(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: _argmin(Tensor self, int64_t dim, bool keepdim=false) -> Tensor
device_guard: false
- func: asin(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: asin_(Tensor self) -> Tensor
CUDA: _asin_out_cuda
- func: atan(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: atan_(Tensor self) -> Tensor
CUDA: _atan_out_cuda
- func: baddbmm(Tensor self, Tensor batch1, Tensor batch2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: baddbmm_cpu
variants: function, method
- func: bilinear(Tensor input1, Tensor input2, Tensor weight, Tensor? bias) -> Tensor
+ matches_jit_signature: True
- func: binary_cross_entropy_with_logits(Tensor self, Tensor target, Tensor? weight, Tensor? pos_weight, int64_t reduction) -> Tensor
variants: function
- func: blackman_window(int64_t window_length, bool periodic, TensorOptions options={}) -> Tensor
- func: bmm(Tensor self, Tensor mat2) -> Tensor
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: bmm_cpu
- func: cat_out(Tensor result, TensorList tensors, int64_t dim=0) -> Tensor
- func: ceil(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: ceil_(Tensor self) -> Tensor
device_guard: false
- func: clamp(Tensor self, Scalar? min=None, Scalar? max=None) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: clamp_(Tensor self, Scalar? min=None, Scalar? max=None) -> Tensor
CUDA: _clamp_out_cuda
- func: clamp_max(Tensor self, Scalar max) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: clamp_max_(Tensor self, Scalar max) -> Tensor
CUDA: _clamp_max_out_cuda
- func: clamp_min(Tensor self, Scalar min) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: clamp_min_(Tensor self, Scalar min) -> Tensor
CUDA: _clamp_min_out_cuda
- func: cudnn_is_acceptable(Tensor self) -> bool
+ matches_jit_signature: True
device_guard: false
- func: constant_pad_nd(Tensor self, IntList pad, Scalar value=0) -> Tensor
variants: function
- func: contiguous(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method
- func: convolution(Tensor input, Tensor weight, Tensor? bias, IntList stride, IntList padding, IntList dilation, bool transposed, IntList output_padding, int64_t groups) -> Tensor
CPU: _copy_same_type__cpu
- func: cos(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: cos_(Tensor self) -> Tensor
CUDA: _cos_out_cuda
- func: cosh(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: cosh_(Tensor self) -> Tensor
CUDA: cudnn_convolution_backward
- func: cudnn_convolution_backward_bias(Tensor grad_output) -> Tensor
+ matches_jit_signature: True
dispatch:
CUDA: cudnn_convolution_backward_bias
CUDA: cudnn_convolution_transpose_backward
- func: cudnn_convolution_transpose_backward_bias(Tensor grad_output) -> Tensor
+ matches_jit_signature: True
dispatch:
CUDA: cudnn_convolution_backward_bias
CUDA: ctc_loss_backward_gpu
- func: det(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: diag_embed(Tensor self, int64_t offset=0, int64_t dim1=-2, int64_t dim2=-1) -> Tensor
variants: function, method
- func: div(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: div_(Tensor self, Tensor other) -> Tensor
# For C++ only, until we have conversion from C++ numbers to Tensor
- func: div(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: div_(Tensor self, Scalar other) -> Tensor
variants: method
- func: dot(Tensor self, Tensor tensor) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: dot_out(Tensor result, Tensor self, Tensor tensor) -> Tensor
device_guard: False
- func: empty_like(Tensor self) -> Tensor
+ matches_jit_signature: True
device_guard: False
- func: empty_like(Tensor self, *, TensorOptions options) -> Tensor
CUDA: empty_strided_cuda
- func: erf(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: erf_(Tensor self) -> Tensor
CUDA: _erf_out_cuda
- func: erfc(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: erfc_(Tensor self) -> Tensor
CUDA: _erfc_out_cuda
- func: exp(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: exp_(Tensor self) -> Tensor
CUDA: _exp_out_cuda
- func: expm1(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: expm1_(Tensor self) -> Tensor
device_guard: false
- func: expand_as(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method # This is method-only to match the previous tensor API. In the future we could make this a function too.
device_guard: false
variants: function, method
- func: floor(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: floor_(Tensor self) -> Tensor
- func: full_out(Tensor result, IntList size, Scalar fill_value) -> Tensor
- func: full_like(Tensor self, Scalar fill_value) -> Tensor
+ matches_jit_signature: True
- func: full_like(Tensor self, Scalar fill_value, *, TensorOptions options) -> Tensor
- func: hinge_embedding_loss(Tensor self, Tensor target, double margin=1.0, int64_t reduction=Reduction::Mean) -> Tensor
- func: ger(Tensor self, Tensor vec2) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: ger_out(Tensor result, Tensor self, Tensor vec2) -> Tensor
- func: gesv(Tensor self, Tensor A) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: function, method
- func: gesv_out(Tensor solution, Tensor lu, Tensor self, Tensor A) -> (Tensor, Tensor)
# gesv handles broadcasting of arbitrary batch dims while _gesv_helper does not.
- func: _gesv_helper(Tensor self, Tensor A) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: function
dispatch:
CPU: _gesv_helper_cpu
variants: function
- func: inverse(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: inverse_out(Tensor result, Tensor self) -> Tensor
- func: _inverse_helper(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function
dispatch:
CPU: _inverse_helper_cpu
variants: function, method
- func: isnan(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function
device_guard: false
- func: is_distributed(Tensor self) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: is_floating_point(Tensor self) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: is_complex(Tensor self) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: is_nonzero(Tensor self) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: is_same_size(Tensor self, Tensor other) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: is_signed(Tensor self) -> bool
+ matches_jit_signature: True
variants: function, method
device_guard: false
- func: linear(Tensor input, Tensor weight, Tensor? bias={}) -> Tensor
- func: fbgemm_linear_int8_weight(Tensor input, Tensor weight, Tensor packed, Tensor col_offsets, Scalar weight_scale, Scalar weight_zero_point, Tensor bias) -> Tensor
+ matches_jit_signature: True
- func: fbgemm_linear_quantize_weight(Tensor input) -> (Tensor, Tensor, double, int64_t)
- func: fbgemm_pack_quantized_matrix(Tensor input, int64_t K, int64_t N) -> Tensor
- func: fbgemm_is_cpu_supported() -> bool
+ matches_jit_signature: True
- func: linspace(Scalar start, Scalar end, int64_t steps=100, TensorOptions options={}) -> Tensor
CUDA: linspace_cuda_out
- func: log(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: log_(Tensor self) -> Tensor
CUDA: _log_out_cuda
- func: log10(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: log10_(Tensor self) -> Tensor
CUDA: _log10_out_cuda
- func: log1p(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: log1p_(Tensor self) -> Tensor
SparseCUDA: log1p_out_sparse
- func: log2(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: log2_(Tensor self) -> Tensor
CUDA: _log2_out_cuda
- func: logdet(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: logspace(Scalar start, Scalar end, int64_t steps=100, TensorOptions options={}) -> Tensor
- func: margin_ranking_loss(Tensor input1, Tensor input2, Tensor target, double margin=0.0, int64_t reduction=Reduction::Mean) -> Tensor
- func: matmul(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: matmul_out(Tensor result, Tensor self, Tensor other) -> Tensor
# FIXME: These could be combined as optional<ScalarType> but for https://github.com/pytorch/pytorch/issues/6593.
- func: mean(Tensor self, *, ScalarType dtype) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mean(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mean(Tensor self, IntList[1] dim, bool keepdim, *, ScalarType dtype) -> Tensor
CUDA: miopen_convolution_backward
- func: miopen_convolution_backward_bias(Tensor grad_output) -> Tensor
+ matches_jit_signature: True
dispatch:
CUDA: miopen_convolution_backward_bias
CUDA: miopen_convolution_transpose_backward_weight
- func: mm(Tensor self, Tensor mat2) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mm_out(Tensor result, Tensor self, Tensor mat2) -> Tensor
- func: _sparse_mm(Tensor sparse, Tensor dense) -> Tensor
+ matches_jit_signature: True
- func: mode(Tensor self, int64_t dim=-1, bool keepdim=false) -> (Tensor, Tensor)
variants: function, method
- func: mode_out(Tensor values, Tensor indices, Tensor self, int64_t dim=-1, bool keepdim=false) -> (Tensor, Tensor)
- func: mul(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mul_(Tensor self, Tensor other) -> Tensor
# For C++ only, until we have conversion from C++ numbers to Tensor
- func: mul(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mul_(Tensor self, Scalar other) -> Tensor
variants: method
- func: mv(Tensor self, Tensor vec) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: mv_out(Tensor result, Tensor self, Tensor vec) -> Tensor
- func: ones_out(Tensor result, IntList size) -> Tensor
- func: ones_like(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: ones_like(Tensor self, *, TensorOptions options) -> Tensor
- func: pixel_shuffle(Tensor self, int64_t upscale_factor) -> Tensor
- func: pin_memory(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: pinverse(Tensor self, double rcond=1e-15) -> Tensor
- func: rand_out(Tensor result, IntList size, *, Generator* generator) -> Tensor
- func: rand_like(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: rand_like(Tensor self, *, TensorOptions options) -> Tensor
- func: randn_out(Tensor result, IntList size, *, Generator* generator) -> Tensor
- func: randn_like(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: randn_like(Tensor self, *, TensorOptions options) -> Tensor
device_guard: false
- func: reshape_as(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method
device_guard: false
CUDA: RoiPooling2d_backward_cuda
- func: round(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: round_(Tensor self) -> Tensor
- func: rrelu_(Tensor self, Scalar lower=0.125, Scalar upper=0.3333333333333333, bool training=false, Generator* generator=nullptr) -> Tensor
- func: relu(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: relu_(Tensor self) -> Tensor
variants: function, method
- func: prelu(Tensor self, Tensor weight) -> Tensor
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: prelu_cpu
CUDA: prelu_cuda
- func: prelu_backward(Tensor grad_output, Tensor self, Tensor weight) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: prelu_backward_cpu
CUDA: prelu_backward_cuda
- func: hardshrink(Tensor self, Scalar lambd=0.5) -> Tensor
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: hardshrink_cpu
CUDA: hardshrink_cuda
- func: hardshrink_backward(Tensor grad_out, Tensor self, Scalar lambd) -> Tensor
+ matches_jit_signature: True
variants: function, method
dispatch:
CPU: hardshrink_backward_cpu
CUDA: hardshrink_backward_cuda
- func: rsqrt(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: rsqrt_(Tensor self) -> Tensor
device_guard: false
- func: selu(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: selu_(Tensor self) -> Tensor
- func: celu(Tensor self, Scalar alpha=1.0) -> Tensor
+ matches_jit_signature: True
- func: celu_(Tensor self, Scalar alpha=1.0) -> Tensor
- func: sigmoid(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sigmoid_(Tensor self) -> Tensor
CUDA: _sigmoid_out_cuda
- func: sin(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sin_(Tensor self) -> Tensor
CUDA: _sin_out_cuda
- func: sinh(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sinh_(Tensor self) -> Tensor
CUDA: _sinh_out_cuda
- func: detach(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: detach_(Tensor self) -> Tensor
device_guard: false
- func: slogdet(Tensor self) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: function, method
- func: smm(Tensor self, Tensor mat2) -> Tensor
+ matches_jit_signature: True
variants: function, method
# FIXME: These could be combined as optional<ScalarType> but for https://github.com/pytorch/pytorch/issues/6593.
device_guard: false
- func: sspaddmm(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sspaddmm_out(Tensor result, Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
# FIXME: These could be combined as optional<ScalarType> but for https://github.com/pytorch/pytorch/issues/6593.
- func: sum(Tensor self, *, ScalarType dtype) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sum(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sum(Tensor self, IntList[1] dim, bool keepdim, *, ScalarType dtype) -> Tensor
device_guard: false
- func: sqrt(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sqrt_(Tensor self) -> Tensor
# FIXME: These could be combined as optional<ScalarType> but for https://github.com/pytorch/pytorch/issues/6593.
- func: prod(Tensor self, *, ScalarType dtype) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: prod(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: prod(Tensor self, int64_t dim, bool keepdim, *, ScalarType dtype) -> Tensor
variants: method
- func: tan(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: tan_(Tensor self) -> Tensor
CUDA: _tan_out_cuda
- func: tanh(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: tanh_(Tensor self) -> Tensor
# TODO: namespace threshold in 'nn'
- func: threshold(Tensor self, Scalar threshold, Scalar value) -> Tensor
+ matches_jit_signature: True
variants: function
- func: threshold_(Tensor self, Scalar threshold, Scalar value) -> Tensor
- func: threshold_out(Tensor result, Tensor self, Scalar threshold, Scalar value) -> Tensor
- func: threshold_backward(Tensor grad_output, Tensor self, Scalar threshold) -> Tensor
+ matches_jit_signature: True
variants: function
- func: transpose(Tensor self, int64_t dim0, int64_t dim1) -> Tensor
- func: triplet_margin_loss(Tensor anchor, Tensor positive, Tensor negative, double margin=1.0, double p=2, double eps=1e-6, bool swap=false, int64_t reduction=Reduction::Mean) -> Tensor
- func: trunc(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: trunc_(Tensor self) -> Tensor
CUDA: _trunc_out_cuda
- func: type_as(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method
- func: _unique(Tensor self, bool sorted=true, bool return_inverse=false) -> (Tensor, Tensor)
- func: var_out(Tensor result, Tensor self, IntList[1] dim, bool unbiased=true, bool keepdim=false) -> Tensor
- func: view_as(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method
device_guard: false
- func: zeros_out(Tensor result, IntList size) -> Tensor
- func: zeros_like(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: zeros_like(Tensor self, *, TensorOptions options) -> Tensor
- func: _standard_gamma_grad(Tensor self, Tensor output) -> Tensor
+ matches_jit_signature: True
variants: function
dispatch:
CPU: _standard_gamma_grad_cpu
# complicated
- func: native_norm(Tensor self, Scalar p=2) -> Tensor
+ matches_jit_signature: True
dispatch:
SparseCPU: norm_sparse
SparseCUDA: norm_sparse
# TODO: reduce signatures down to one when optional args is available
- func: _sparse_sum(Tensor self) -> Tensor
+ matches_jit_signature: True
- func: _sparse_sum(Tensor self, *, ScalarType dtype) -> Tensor
+ matches_jit_signature: True
- func: _sparse_sum(Tensor self, IntList[1] dim) -> Tensor
variants: function, method
- func: norm(Tensor self, Scalar p=2) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: norm(Tensor self, Scalar? p, IntList[1] dim, bool keepdim, *, ScalarType dtype) -> Tensor
- func: norm_out(Tensor result, Tensor self, Scalar? p, IntList[1] dim, bool keepdim=false) -> Tensor
- func: frobenius_norm(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function
- func: frobenius_norm(Tensor self, IntList[1] dim, bool keepdim=false) -> Tensor
variants: function
- func: native_clone(Tensor self) -> Tensor
+ matches_jit_signature: True
dispatch:
SparseCPU: clone_sparse
SparseCUDA: clone_sparse
- func: clone(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: native_resize_as_(Tensor self, Tensor the_template) -> Tensor
SparseCUDA: pow_out_sparse_scalar
- func: native_pow(Tensor self, Scalar exponent) -> Tensor
+ matches_jit_signature: True
dispatch:
SparseCPU: pow_sparse_scalar
SparseCUDA: pow_sparse_scalar
- func: pow_out(Tensor result, Tensor self, Scalar exponent) -> Tensor
- func: pow(Tensor self, Scalar exponent) -> Tensor
+ matches_jit_signature: True
variants: function, method
variants: method, function
- func: sub_out(Tensor result, Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
- func: sub(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sub_(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
# For C++ only, until we have conversion from C++ numbers to Tensor
- func: sub(Tensor self, Scalar other, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: sub_(Tensor self, Scalar other, Scalar alpha=1) -> Tensor
variants: method
- func: rsub(Tensor self, Tensor other, *, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function
# For C++ only, until we have conversion from C++ numbers to Tensor
- func: rsub(Tensor self, Scalar other, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function
- func: s_native_addmm_out(Tensor result, Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
CUDA: s_addmm_out_sparse_dense_cuda
- func: s_native_addmm(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
dispatch:
CPU: s_addmm_sparse_dense_cpu
CUDA: s_addmm_sparse_dense_cuda
CUDA: s_addmm_sparse_dense_cuda_
- func: _sparse_addmm(Tensor self, Tensor sparse, Tensor dense, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
- func: addmm_out(Tensor result, Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
- func: addmm(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: function, method
- func: addmm_(Tensor self, Tensor mat1, Tensor mat2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
- func: to_dense(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method
dispatch:
SparseCPU: sparse_to_dense
- func: coalesce(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method
dispatch:
SparseCPU: coalesce_sparse_cpu
- func: is_coalesced(Tensor self) -> bool
+ matches_jit_signature: True
variants: method
dispatch:
SparseCPU: is_coalesced_sparse
requires_tensor: True
- func: hspmm(Tensor mat1, Tensor mat2) -> Tensor
+ matches_jit_signature: True
dispatch:
SparseCPU: hspmm_sparse_cpu
SparseCUDA: hspmm_sparse_cuda
CUDA: dense_to_sparse
- func: to_sparse(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method
dispatch:
CPU: dense_to_sparse
variants: function
- func: item(Tensor self) -> Scalar
+ matches_jit_signature: True
variants: method
# NB: Does NOT check precondition that numel == 1
# WARNING: Use of cpu_half here is generally not supported; please
# don't use it.
- func: _local_scalar_dense(Tensor self) -> Scalar
+ matches_jit_signature: True
cpu_half: True
dispatch:
CPU: _local_scalar_dense_cpu
CUDA: _thnn_fused_lstm_cell_cuda
- func: _thnn_fused_lstm_cell_backward(Tensor? grad_hy, Tensor? grad_cy, Tensor cx, Tensor cy, Tensor workspace, bool has_bias) -> (Tensor, Tensor, Tensor, Tensor, Tensor)
+ matches_jit_signature: True
dispatch:
CUDA: _thnn_fused_lstm_cell_backward_cuda
CUDA: _thnn_fused_gru_cell_cuda
- func: _thnn_fused_gru_cell_backward(Tensor grad_hy, Tensor workspace, bool has_bias) -> (Tensor, Tensor, Tensor, Tensor, Tensor)
+ matches_jit_signature: True
dispatch:
CUDA: _thnn_fused_gru_cell_backward_cuda
# PackedSequence utilities
- func: _pack_padded_sequence(Tensor input, Tensor lengths, bool batch_first) -> (Tensor, Tensor)
+ matches_jit_signature: True
- func: _pack_padded_sequence_backward(Tensor grad, IntList input_size, Tensor batch_sizes, bool batch_first) -> Tensor
device_guard: false
- func: is_set_to(Tensor self, Tensor tensor) -> bool
+ matches_jit_signature: True
variants: method
device_guard: false
variants: method
- func: __and__(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __and__(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __iand__(Tensor self, Scalar other) -> Tensor
variants: method
- func: __or__(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __or__(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __ior__(Tensor self, Scalar other) -> Tensor
variants: method
- func: __xor__(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __xor__(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __ixor__(Tensor self, Scalar other) -> Tensor
variants: method
- func: __lshift__(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __lshift__(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __ilshift__(Tensor self, Scalar other) -> Tensor
variants: method
- func: __rshift__(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __rshift__(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: __irshift__(Tensor self, Scalar other) -> Tensor
- func: addbmm_out(Tensor result, Tensor self, Tensor batch1, Tensor batch2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
- func: addbmm(Tensor self, Tensor batch1, Tensor batch2, *, Scalar beta=1, Scalar alpha=1) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: addcmul_(Tensor self, Tensor tensor1, Tensor tensor2, *, Scalar value=1) -> Tensor
CUDA: triu_indices_cuda
- func: trace(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: ne_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: ne(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: ne_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: ne(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: eq_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: eq(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: eq_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: eq(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: ge_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: ge(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: ge_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: ge(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: le_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: le(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: le_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: le(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: gt_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: gt(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: gt_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: gt(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: lt_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: lt(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: lt_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: lt(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: take_out(Tensor result, Tensor self, Tensor index) -> Tensor
- func: take(Tensor self, Tensor index) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: index_select_out(Tensor result, Tensor self, int64_t dim, Tensor index) -> Tensor
- func: masked_select_out(Tensor result, Tensor self, Tensor mask) -> Tensor
- func: masked_select(Tensor self, Tensor mask) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: nonzero_out(Tensor result, Tensor self) -> Tensor
- func: nonzero(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: gather_out(Tensor result, Tensor self, int64_t dim, Tensor index) -> Tensor
- func: addcmul_out(Tensor result, Tensor self, Tensor tensor1, Tensor tensor2, *, Scalar value=1) -> Tensor
- func: addcmul(Tensor self, Tensor tensor1, Tensor tensor2, *, Scalar value=1) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: addcdiv_out(Tensor result, Tensor self, Tensor tensor1, Tensor tensor2, *, Scalar value=1) -> Tensor
- func: addcdiv(Tensor self, Tensor tensor1, Tensor tensor2, *, Scalar value=1) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: gels_out(Tensor X, Tensor qr, Tensor self, Tensor A) -> (Tensor, Tensor)
- func: gels(Tensor self, Tensor A) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: method, function
- func: trtrs_out(Tensor X, Tensor M, Tensor self, Tensor A, bool upper=true, bool transpose=false, bool unitriangular=false) -> (Tensor, Tensor)
variants: method, function
- func: _cholesky_helper(Tensor self, bool upper) -> Tensor
+ matches_jit_signature: True
variants: function
dispatch:
CPU: _cholesky_helper_cpu
variants: method, function
- func: _cholesky_solve_helper(Tensor self, Tensor A, bool upper) -> Tensor
+ matches_jit_signature: True
variants: function
dispatch:
CPU: _cholesky_solve_helper_cpu
- func: qr_out(Tensor Q, Tensor R, Tensor self) -> (Tensor, Tensor)
- func: qr(Tensor self) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: method, function
- func: geqrf_out(Tensor result0, Tensor result1, Tensor self) -> (Tensor, Tensor)
- func: geqrf(Tensor self) -> (Tensor, Tensor)
+ matches_jit_signature: True
variants: method, function
- func: orgqr_out(Tensor result, Tensor self, Tensor input2) -> Tensor
- func: orgqr(Tensor self, Tensor input2) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: ormqr_out(Tensor result, Tensor self, Tensor input2, Tensor input3, bool left=true, bool transpose=false) -> Tensor
- func: btrisolve_out(Tensor result, Tensor self, Tensor LU_data, Tensor LU_pivots) -> Tensor
- func: btrisolve(Tensor self, Tensor LU_data, Tensor LU_pivots) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: multinomial_out(Tensor result, Tensor self, int64_t num_samples, bool replacement=false, *, Generator* generator=nullptr) -> Tensor
- func: lgamma_out(Tensor result, Tensor self) -> Tensor
- func: lgamma(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: digamma_out(Tensor result, Tensor self) -> Tensor
- func: digamma(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: polygamma_out(Tensor result, int64_t n, Tensor self) -> Tensor
- func: erfinv_out(Tensor result, Tensor self) -> Tensor
- func: erfinv(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: frac_out(Tensor result, Tensor self) -> Tensor
- func: frac(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: dist(Tensor self, Tensor other, Scalar p=2) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: reciprocal_out(Tensor result, Tensor self) -> Tensor
- func: reciprocal(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: neg_out(Tensor result, Tensor self) -> Tensor
- func: neg(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: atan2_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: atan2(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: lerp_out(Tensor result, Tensor self, Tensor end, Scalar weight) -> Tensor
- func: lerp(Tensor self, Tensor end, Scalar weight) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: histc_out(Tensor result, Tensor self, int64_t bins=100, Scalar min=0, Scalar max=0) -> Tensor
- func: sign_out(Tensor result, Tensor self) -> Tensor
- func: sign(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: fmod_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: fmod(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: fmod_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: fmod(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: remainder_out(Tensor result, Tensor self, Scalar other) -> Tensor
- func: remainder(Tensor self, Scalar other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: remainder_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: remainder(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: min_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: min(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: min(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: max_out(Tensor result, Tensor self, Tensor other) -> Tensor
- func: max(Tensor self, Tensor other) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: max(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: median(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: sort_out(Tensor values, Tensor indices, Tensor self, int64_t dim=-1, bool descending=false) -> (Tensor, Tensor)
variants: method, function
- func: all(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: any(Tensor self) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: renorm_out(Tensor result, Tensor self, Scalar p, int64_t dim, Scalar maxnorm) -> Tensor
device_guard: false
- func: equal(Tensor self, Tensor other) -> bool
+ matches_jit_signature: True
variants: method, function
- func: pow_out(Tensor result, Tensor self, Tensor exponent) -> Tensor
- func: pow(Tensor self, Tensor exponent) -> Tensor
+ matches_jit_signature: True
variants: method, function
- func: pow_out(Tensor result, Scalar self, Tensor exponent) -> Tensor
- func: pow(Scalar self, Tensor exponent) -> Tensor
+ matches_jit_signature: True
- func: normal_out(Tensor output, Tensor mean, double std=1, *, Generator* generator=nullptr) -> Tensor
- func: _dirichlet_grad_out(Tensor output, Tensor x, Tensor alpha, Tensor total) -> Tensor
- func: _dirichlet_grad(Tensor x, Tensor alpha, Tensor total) -> Tensor
+ matches_jit_signature: True
## NN wrappers
python_module: nn
- func: elu(Tensor self, Scalar alpha=1, Scalar scale=1, Scalar input_scale=1) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: elu_backward_out(Tensor grad_input, Tensor grad_output, Scalar alpha, Scalar scale, Scalar input_scale, Tensor output) -> Tensor
python_module: nn
- func: elu_backward(Tensor grad_output, Scalar alpha, Scalar scale, Scalar input_scale, Tensor output) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: elu_(Tensor self, Scalar alpha=1, Scalar scale=1, Scalar input_scale=1) -> Tensor
python_module: nn
- func: hardtanh(Tensor self, Scalar min_val=-1, Scalar max_val=1) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: hardtanh_backward_out(Tensor grad_input, Tensor grad_output, Tensor self, Scalar min_val, Scalar max_val) -> Tensor
python_module: nn
- func: hardtanh_backward(Tensor grad_output, Tensor self, Scalar min_val, Scalar max_val) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: hardtanh_(Tensor self, Scalar min_val=-1, Scalar max_val=1) -> Tensor
python_module: nn
- func: leaky_relu(Tensor self, Scalar negative_slope=0.01) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: leaky_relu_backward_out(Tensor grad_input, Tensor grad_output, Tensor self, Scalar negative_slope) -> Tensor
python_module: nn
- func: leaky_relu_backward(Tensor grad_output, Tensor self, Scalar negative_slope) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: leaky_relu_(Tensor self, Scalar negative_slope=0.01) -> Tensor
python_module: nn
- func: log_sigmoid(Tensor self) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: log_sigmoid_forward_out(Tensor output, Tensor buffer, Tensor self) -> (Tensor, Tensor)
python_module: nn
- func: log_sigmoid_backward(Tensor grad_output, Tensor self, Tensor buffer) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: rrelu_with_noise_out(Tensor output, Tensor self, Tensor noise, Scalar lower=0.125, Scalar upper=0.3333333333333333, bool training=false, Generator* generator=nullptr) -> Tensor
python_module: nn
- func: rrelu_with_noise_backward(Tensor grad_output, Tensor self, Tensor noise, Scalar lower, Scalar upper, bool training) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: rrelu_with_noise_(Tensor self, Tensor noise, Scalar lower=0.125, Scalar upper=0.3333333333333333, bool training=false, Generator* generator=nullptr) -> Tensor
python_module: nn
- func: softplus(Tensor self, Scalar beta=1, Scalar threshold=20) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: softplus_backward_out(Tensor grad_input, Tensor grad_output, Tensor self, Scalar beta, Scalar threshold, Tensor output) -> Tensor
python_module: nn
- func: softplus_backward(Tensor grad_output, Tensor self, Scalar beta, Scalar threshold, Tensor output) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: softshrink_out(Tensor output, Tensor self, Scalar lambd=0.5) -> Tensor
python_module: nn
- func: softshrink(Tensor self, Scalar lambd=0.5) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: softshrink_backward_out(Tensor grad_input, Tensor grad_output, Tensor self, Scalar lambd) -> Tensor
python_module: nn
- func: softshrink_backward(Tensor grad_output, Tensor self, Scalar lambd) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: adaptive_avg_pool2d_out(Tensor output, Tensor self, IntList[2] output_size) -> Tensor
CUDA: adaptive_avg_pool2d_backward_out_cuda
- func: adaptive_avg_pool2d_backward(Tensor grad_output, Tensor self) -> Tensor
+ matches_jit_signature: True
python_module: nn
dispatch:
CPU: adaptive_avg_pool2d_backward_cpu
python_module: nn
- func: adaptive_avg_pool3d_backward(Tensor grad_output, Tensor self) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: adaptive_max_pool2d_out(Tensor output, Tensor indices, Tensor self, IntList[2] output_size) -> (Tensor output, Tensor indices)
python_module: nn
- func: adaptive_max_pool2d_backward(Tensor grad_output, Tensor self, Tensor indices) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: adaptive_max_pool3d_out(Tensor output, Tensor indices, Tensor self, IntList[3] output_size) -> (Tensor output, Tensor indices)
python_module: nn
- func: adaptive_max_pool3d_backward(Tensor grad_output, Tensor self, Tensor indices) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: avg_pool2d_out(Tensor output, Tensor self, IntList[2] kernel_size, IntList[2] stride={}, IntList[2] padding=0, bool ceil_mode=false, bool count_include_pad=true) -> Tensor
python_module: nn
- func: sigmoid_backward(Tensor grad_output, Tensor output) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: tanh_backward_out(Tensor grad_input, Tensor grad_output, Tensor output) -> Tensor
python_module: nn
- func: tanh_backward(Tensor grad_output, Tensor output) -> Tensor
+ matches_jit_signature: True
python_module: nn
- func: thnn_conv_transpose2d_out(Tensor output, Tensor self, Tensor weight, IntList[2] kernel_size, Tensor? bias={}, IntList[2] stride=1, IntList[2] padding=0, IntList[2] output_padding=0, IntList[2] dilation=1) -> Tensor