infini::ops::PagedAttentionPrefillInfinilm Member List

This is the complete list of members for infini::ops::PagedAttentionPrefillInfinilm, including all inherited members.

active_implementation_indices(Device::Type dev_type)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
active_implementation_indices(Device::Type dev_type)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
block_size_	infini::ops::PagedAttentionPrefillInfinilm	protected
block_table_batch_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
Call(const Handle &handle, const Config &config, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	static
Call(const Tensor tensor, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Call(const TensorLike &tensor, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Call(const Handle &handle, const Config &config, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Call(const Tensor tensor, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Call(const TensorLike &tensor, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
clear_cache()	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
clear_cache()	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
config_ptr_	infini::ops::OperatorBase	protected
device_type_	infini::ops::Operator< PagedAttentionPrefillInfinilm >	protectedstatic
dtype_	infini::ops::PagedAttentionPrefillInfinilm	protected
handle_ptr_	infini::ops::OperatorBase	protected
head_size_	infini::ops::PagedAttentionPrefillInfinilm	protected
implementation_index_	infini::ops::Operator< PagedAttentionPrefillInfinilm >	protectedstatic
index_dtype_	infini::ops::PagedAttentionPrefillInfinilm	protected
IsIndexDtype(DataType dtype)	infini::ops::PagedAttentionPrefillInfinilm	inlineprotectedstatic
k_cache_block_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
k_cache_head_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
k_cache_slot_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
Make(const Config &config, const Tensor tensor, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const Tensor tensor, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const Config &config, const std::vector< Tensor > tensors, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const std::vector< Tensor > tensors, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const Config &config, const Tensor tensor, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const Tensor tensor, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const Config &config, const std::vector< Tensor > tensors, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
Make(const std::vector< Tensor > tensors, Args &&... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inlinestatic
max_num_blocks_per_seq_	infini::ops::PagedAttentionPrefillInfinilm	protected
num_heads_	infini::ops::PagedAttentionPrefillInfinilm	protected
num_kv_heads_	infini::ops::PagedAttentionPrefillInfinilm	protected
num_seqs_	infini::ops::PagedAttentionPrefillInfinilm	protected
operator()(const Tensor q, const Tensor k_cache, const Tensor v_cache, const Tensor block_tables, const Tensor seq_lens, const Tensor cum_seq_lens_q, std::optional< Tensor > alibi_slopes, float scale, Tensor out) const =0	infini::ops::PagedAttentionPrefillInfinilm	pure virtual
Operator< PagedAttentionPrefillInfinilm >::operator()(const Handle &handle, const Args &... args)	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inline
Operator< PagedAttentionPrefillInfinilm >::operator()(const Args &... args) const	infini::ops::Operator< PagedAttentionPrefillInfinilm >	inline
out_head_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
out_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
PagedAttentionPrefillInfinilm(const Tensor q, const Tensor k_cache, const Tensor v_cache, const Tensor block_tables, const Tensor seq_lens, const Tensor cum_seq_lens_q, std::optional< Tensor > alibi_slopes, float scale, Tensor out)	infini::ops::PagedAttentionPrefillInfinilm	inline
q_head_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
q_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
scale_	infini::ops::PagedAttentionPrefillInfinilm	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
total_q_tokens_	infini::ops::PagedAttentionPrefillInfinilm	protected
v_cache_block_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
v_cache_head_stride_	infini::ops::PagedAttentionPrefillInfinilm	protected
v_cache_slot_stride_	infini::ops::PagedAttentionPrefillInfinilm	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