diff options
| author | Mitja Felicijan <mitja.felicijan@gmail.com> | 2026-02-12 20:57:17 +0100 |
|---|---|---|
| committer | Mitja Felicijan <mitja.felicijan@gmail.com> | 2026-02-12 20:57:17 +0100 |
| commit | b333b06772c89d96aacb5490d6a219fba7c09cc6 (patch) | |
| tree | 211df60083a5946baa2ed61d33d8121b7e251b06 /llama.cpp/src/llama-memory-hybrid.h | |
| download | llmnpc-b333b06772c89d96aacb5490d6a219fba7c09cc6.tar.gz | |
Engage!
Diffstat (limited to 'llama.cpp/src/llama-memory-hybrid.h')
| -rw-r--r-- | llama.cpp/src/llama-memory-hybrid.h | 139 |
1 files changed, 139 insertions, 0 deletions
diff --git a/llama.cpp/src/llama-memory-hybrid.h b/llama.cpp/src/llama-memory-hybrid.h new file mode 100644 index 0000000..558cafd --- /dev/null +++ b/llama.cpp/src/llama-memory-hybrid.h @@ -0,0 +1,139 @@ +#pragma once + +#include "llama-batch.h" +#include "llama-graph.h" +#include "llama-kv-cache.h" +#include "llama-memory.h" +#include "llama-memory-recurrent.h" + +#include <memory> +#include <vector> + +// +// llama_memory_hybrid +// + +// utilizes instances of llama_memory_recurrent and llama_kv_cache to +// support models where each layer may be either attention-based or recurrent + +class llama_memory_hybrid : public llama_memory_i { +public: + llama_memory_hybrid( + const llama_model & model, + /* attn */ + ggml_type type_k, + ggml_type type_v, + bool v_trans, + uint32_t kv_size, + uint32_t n_pad, + uint32_t n_swa, + llama_swa_type swa_type, + /* recurrent */ + ggml_type type_r, + ggml_type type_s, + uint32_t rs_size, + /* common */ + uint32_t n_seq_max, + bool offload, + bool unified, + /* layer filters */ + const layer_filter_cb & filter_attn = nullptr, + const layer_filter_cb & filter_recr = nullptr); + + ~llama_memory_hybrid() = default; + + // + // llama_memory_i + // + + llama_memory_context_ptr init_batch( + llama_batch_allocr & balloc, + uint32_t n_ubatch, + bool embd_all) override; + + llama_memory_context_ptr init_full() override; + + llama_memory_context_ptr init_update(llama_context * lctx, bool optimize) override; + + bool get_can_shift() const override; + + void clear(bool data) override; + + bool seq_rm (llama_seq_id seq_id, llama_pos p0, llama_pos p1) override; + void seq_cp (llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) override; + void seq_keep(llama_seq_id seq_id) override; + void seq_add (llama_seq_id seq_id, llama_pos p0, llama_pos p1, llama_pos shift) override; + void seq_div (llama_seq_id seq_id, llama_pos p0, llama_pos p1, int d) override; + + llama_pos seq_pos_min(llama_seq_id seq_id) const override; + llama_pos seq_pos_max(llama_seq_id seq_id) const override; + + std::map<ggml_backend_buffer_type_t, size_t> memory_breakdown() const override; + + // state write/load + + void state_write(llama_io_write_i & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) const override; + void state_read (llama_io_read_i & io, llama_seq_id seq_id = -1, llama_state_seq_flags flags = 0) override; + + // + // llama_memory_hybrid specific API + // + + llama_kv_cache * get_mem_attn() const; + llama_memory_recurrent * get_mem_recr() const; + +private: + const llama_hparams & hparams; + + const std::unique_ptr<llama_kv_cache> mem_attn; + const std::unique_ptr<llama_memory_recurrent> mem_recr; +}; + +class llama_memory_hybrid_context : public llama_memory_context_i { +public: + using slot_info_vec_t = llama_kv_cache::slot_info_vec_t; + + // init failure + explicit llama_memory_hybrid_context(llama_memory_status status); + + // init full + explicit llama_memory_hybrid_context(llama_memory_hybrid * mem); + + // init update + explicit llama_memory_hybrid_context( + llama_memory_hybrid * mem, + llama_context * lctx, + bool optimize); + + // init success + llama_memory_hybrid_context( + llama_memory_hybrid * mem, + slot_info_vec_t sinfos_attn, + std::vector<llama_ubatch> ubatches); + + ~llama_memory_hybrid_context() = default; + + bool next() override; + bool apply() override; + + llama_memory_status get_status() const override; + const llama_ubatch & get_ubatch() const override; + + // + // llama_memory_hybrid_context + // + + const llama_kv_cache_context * get_attn() const; + const llama_memory_recurrent_context * get_recr() const; + +private: + // the index of the next ubatch to process + size_t i_next = 0; + + std::vector<llama_ubatch> ubatches; + + const llama_memory_context_ptr ctx_attn; + const llama_memory_context_ptr ctx_recr; + + const llama_memory_status status; +}; |
