| active_implementation_indices(Device::Type dev_type) | infini::ops::Operator< FlashAttention > | inlinestatic |
| active_implementation_indices(Device::Type dev_type) | infini::ops::Operator< FlashAttention > | inlinestatic |
| block_size_ | infini::ops::FlashAttention | protected |
| Call(const Handle &handle, const Config &config, const Args &... args) | infini::ops::Operator< FlashAttention > | static |
| Call(const Tensor tensor, const Args &... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Call(const TensorLike &tensor, const Args &... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Call(const Handle &handle, const Config &config, const Args &... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Call(const Tensor tensor, const Args &... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Call(const TensorLike &tensor, const Args &... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| causal_ | infini::ops::FlashAttention | protected |
| clear_cache() | infini::ops::Operator< FlashAttention > | inlinestatic |
| clear_cache() | infini::ops::Operator< FlashAttention > | inlinestatic |
| config_ptr_ | infini::ops::OperatorBase | protected |
| device_type_ | infini::ops::Operator< FlashAttention > | protectedstatic |
| dtype_ | infini::ops::FlashAttention | protected |
| FlashAttention(const Tensor query, const Tensor key, const Tensor value, std::optional< Tensor > cu_seqlens_q, std::optional< Tensor > cu_seqlens_kv, std::optional< Tensor > block_table, int64_t num_heads, int64_t num_kv_heads, int64_t head_size, double scale, bool causal, int64_t window_left, int64_t window_right, int64_t block_size, Tensor output) | infini::ops::FlashAttention | inline |
| handle_ptr_ | infini::ops::OperatorBase | protected |
| has_block_table_ | infini::ops::FlashAttention | protected |
| has_cu_seqlens_kv_ | infini::ops::FlashAttention | protected |
| has_cu_seqlens_q_ | infini::ops::FlashAttention | protected |
| head_size_ | infini::ops::FlashAttention | protected |
| implementation_index_ | infini::ops::Operator< FlashAttention > | protectedstatic |
| key_shape_ | infini::ops::FlashAttention | protected |
| key_strides_ | infini::ops::FlashAttention | protected |
| Make(const Config &config, const Tensor tensor, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const Tensor tensor, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const Config &config, const std::vector< Tensor > tensors, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const std::vector< Tensor > tensors, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const Config &config, const Tensor tensor, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const Tensor tensor, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const Config &config, const std::vector< Tensor > tensors, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| Make(const std::vector< Tensor > tensors, Args &&... args) | infini::ops::Operator< FlashAttention > | inlinestatic |
| num_heads_ | infini::ops::FlashAttention | protected |
| num_kv_heads_ | infini::ops::FlashAttention | protected |
| num_tokens_ | infini::ops::FlashAttention | protected |
| operator()(const Tensor query, const Tensor key, const Tensor value, std::optional< Tensor > cu_seqlens_q, std::optional< Tensor > cu_seqlens_kv, std::optional< Tensor > block_table, int64_t num_heads, int64_t num_kv_heads, int64_t head_size, double scale, bool causal, int64_t window_left, int64_t window_right, int64_t block_size, Tensor output) const =0 | infini::ops::FlashAttention | pure virtual |
| Operator< FlashAttention >::operator()(const Handle &handle, const Args &... args) | infini::ops::Operator< FlashAttention > | inline |
| Operator< FlashAttention >::operator()(const Args &... args) const | infini::ops::Operator< FlashAttention > | inline |
| output_shape_ | infini::ops::FlashAttention | protected |
| output_strides_ | infini::ops::FlashAttention | protected |
| query_shape_ | infini::ops::FlashAttention | protected |
| query_strides_ | infini::ops::FlashAttention | protected |
| scale_ | infini::ops::FlashAttention | protected |
| set_config(const Config &config) | infini::ops::OperatorBase | inline |
| set_config(const Config &config) | infini::ops::OperatorBase | inline |
| set_handle(const Handle &handle) | infini::ops::OperatorBase | inline |
| set_handle(const Handle &handle) | infini::ops::OperatorBase | inline |
| set_stream(void *stream) | infini::ops::OperatorBase | inline |
| set_stream(void *stream) | infini::ops::OperatorBase | inline |
| set_workspace(void *workspace) | infini::ops::OperatorBase | inline |
| set_workspace(void *workspace) | infini::ops::OperatorBase | inline |
| set_workspace_size_in_bytes(std::size_t workspace_size_in_bytes) | infini::ops::OperatorBase | inline |
| set_workspace_size_in_bytes(std::size_t workspace_size_in_bytes) | infini::ops::OperatorBase | inline |
| stream_ | infini::ops::OperatorBase | protected |
| value_shape_ | infini::ops::FlashAttention | protected |
| value_strides_ | infini::ops::FlashAttention | protected |
| window_left_ | infini::ops::FlashAttention | protected |
| window_right_ | infini::ops::FlashAttention | protected |
| workspace_ | infini::ops::OperatorBase | protected |
| workspace_size_in_bytes() const | infini::ops::OperatorBase | inlinevirtual |
| workspace_size_in_bytes() const | infini::ops::OperatorBase | inlinevirtual |
| workspace_size_in_bytes_ | infini::ops::OperatorBase | protected |
| ~OperatorBase()=default | infini::ops::OperatorBase | virtual |
| ~OperatorBase()=default | infini::ops::OperatorBase | virtual |
1.9.8