Function at::_scaled_dot_product_efficient_attention_backward¶

Function Documentation¶

inline ::std::tuple<at::Tensor, at::Tensor, at::Tensor, at::Tensor> at::_scaled_dot_product_efficient_attention_backward(const at::Tensor &grad_out_, const at::Tensor &query, const at::Tensor &key, const at::Tensor &value, const at::Tensor &attn_bias, const at::Tensor &out, const at::Tensor &logsumexp, const at::Tensor &philox_seed, const at::Tensor &philox_offset, double dropout_p, ::std::array<bool, 4> grad_input_mask, bool is_causal = false, ::std::optional<double> scale = ::std::nullopt)¶