tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-14 22:58:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

context : remove redundant virtual, protected -> private

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-02-23 20:02:11 +02:00
parent 6378112cb5
commit 0699a44c83
4 changed files with 79 additions and 66 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llama-context.cpp
View File

@ -1842,6 +1842,14 @@ ggml_tensor * llama_context::build_attn(
return cur; return cur;
} }
ggml_tensor * llama_context::build_inp_self_k_shift(
ggml_context * ctx0) {
GGML_UNUSED(ctx0);
LLAMA_LOG_ERROR("%s: not implemented\n", __func__);
return nullptr;
}
void llama_context::build_kv_self_shift( void llama_context::build_kv_self_shift(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf) { ggml_cgraph * gf) {

126
src/llama-context.h
View File

@ -140,6 +140,7 @@ protected:
virtual void input_set(const llama_ubatch & ubatch); virtual void input_set(const llama_ubatch & ubatch);
private:
struct { struct {
// base input tensors // base input tensors
ggml_tensor * tokens; // I32 [n_batch] ggml_tensor * tokens; // I32 [n_batch]
@ -155,6 +156,7 @@ protected:
ggml_tensor * kq_mask_cnv; // [n_tokens, n_batch] ggml_tensor * kq_mask_cnv; // [n_tokens, n_batch]
} inp; } inp;
protected:
// //
// output // output
// //
@ -192,71 +194,71 @@ public:
// graph build // graph build
// //
virtual void build_cb( void build_cb(
ggml_tensor * cur, ggml_tensor * cur,
const char * name, const char * name,
const llama_ubatch & ubatch, const llama_ubatch & ubatch,
int il) override; int il) override;
// apply control vector for layer il // apply control vector for layer il
virtual ggml_tensor * build_cvec( ggml_tensor * build_cvec(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * cur, ggml_tensor * cur,
int il) override; int il) override;
// do mat_mul, while optionally apply lora // do mat_mul, while optionally apply lora
virtual ggml_tensor * build_lora_mm( ggml_tensor * build_lora_mm(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * w, ggml_tensor * w,
ggml_tensor * cur) override; ggml_tensor * cur) override;
// do mat_mul_id, while optionally apply lora // do mat_mul_id, while optionally apply lora
virtual ggml_tensor * build_lora_mm_id( ggml_tensor * build_lora_mm_id(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * w, // struct ggml_tensor * as ggml_tensor * w, // struct ggml_tensor * as
ggml_tensor * cur, // struct ggml_tensor * b ggml_tensor * cur, // struct ggml_tensor * b
ggml_tensor * ids) override; ggml_tensor * ids) override;
virtual ggml_tensor * build_rope_factors(int il) override; ggml_tensor * build_rope_factors(int il) override;
virtual ggml_tensor * build_rope_shift( ggml_tensor * build_rope_shift(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * cur, ggml_tensor * cur,
ggml_tensor * shift, ggml_tensor * shift,
ggml_tensor * factors, ggml_tensor * factors,
ggml_backend_buffer * bbuf) override; ggml_backend_buffer * bbuf) override;
virtual ggml_tensor * build_inp_embd( ggml_tensor * build_inp_embd(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * tok_embd, ggml_tensor * tok_embd,
const llama_ubatch & ubatch) override; const llama_ubatch & ubatch) override;
virtual ggml_tensor * build_inp_pos( ggml_tensor * build_inp_pos(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
virtual ggml_tensor * build_inp_pos_bucket( ggml_tensor * build_inp_pos_bucket(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
virtual ggml_tensor * build_inp_out_ids( ggml_tensor * build_inp_out_ids(
ggml_context * ctx0) override; ggml_context * ctx0) override;
virtual ggml_tensor * build_inp_mean( ggml_tensor * build_inp_mean(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
virtual ggml_tensor * build_inp_cls( ggml_tensor * build_inp_cls(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
virtual void build_attn_inp( void build_attn_inp(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens, int32_t n_tokens,
bool causal, bool causal,
bool swa) override; bool swa) override;
virtual ggml_tensor * build_attn( ggml_tensor * build_attn(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * wo, ggml_tensor * wo,
@ -270,6 +272,9 @@ public:
int il) override; int il) override;
protected: protected:
virtual ggml_tensor * build_inp_self_k_shift(
ggml_context * ctx0);
virtual void build_kv_self_shift( virtual void build_kv_self_shift(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf); ggml_cgraph * gf);
@ -288,6 +293,7 @@ public:
virtual void perf_reset(); virtual void perf_reset();
protected: protected:
// TODO: become private
mutable int64_t t_start_us = 0; mutable int64_t t_start_us = 0;
mutable int64_t t_load_us = 0; mutable int64_t t_load_us = 0;
mutable int64_t t_p_eval_us = 0; mutable int64_t t_p_eval_us = 0;
@ -346,6 +352,7 @@ protected:
// //
// members // members
// //
// TODO: become private / move to llama_graph_i
const llama_model & model; const llama_model & model;
@ -412,24 +419,25 @@ public:
virtual ~llama_context_kv_self(); virtual ~llama_context_kv_self();
protected: protected:
virtual void reserve() override; void reserve() override;
public: public:
virtual llama_kv_cache * get_kv_self() override; llama_kv_cache * get_kv_self() override;
virtual const llama_kv_cache * get_kv_self() const override; const llama_kv_cache * get_kv_self() const override;
virtual void kv_self_update() override; void kv_self_update() override;
virtual int encode(llama_batch & inp_batch) override; int encode(llama_batch & inp_batch) override;
virtual int decode(llama_batch & inp_batch) override; int decode(llama_batch & inp_batch) override;
protected: protected:
// //
// input // input
// //
virtual void input_set(const llama_ubatch & ubatch) override; void input_set(const llama_ubatch & ubatch) override;
private:
struct { struct {
ggml_tensor * self_pos_bucket; // I32 [n_kv, n_batch] ggml_tensor * self_pos_bucket; // I32 [n_kv, n_batch]
ggml_tensor * self_kq_mask; // F32 [n_kv, n_batch] ggml_tensor * self_kq_mask; // F32 [n_kv, n_batch]
@ -443,26 +451,24 @@ protected:
// graph // graph
// //
virtual ggml_cgraph * graph_init() override; ggml_cgraph * graph_init() override;
public: public:
// //
// graph build // graph build
// //
virtual ggml_tensor * build_inp_self_k_shift(ggml_context * ctx0) override; ggml_tensor * build_inp_pos_bucket(
virtual ggml_tensor * build_inp_pos_bucket(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
virtual void build_attn_inp( void build_attn_inp(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens, int32_t n_tokens,
bool causal, bool causal,
bool swa) override; bool swa) override;
virtual ggml_tensor * build_attn( ggml_tensor * build_attn(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * wo, ggml_tensor * wo,
@ -476,16 +482,22 @@ public:
int il) override; int il) override;
protected: protected:
virtual void build_kv_self_shift( ggml_tensor * build_inp_self_k_shift(ggml_context * ctx0) override;
void build_kv_self_shift(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf) override; ggml_cgraph * gf) override;
// find holes from the beginning of the KV cache and fill them by moving data from the end of the cache // find holes from the beginning of the KV cache and fill them by moving data from the end of the cache
virtual void build_kv_self_defrag( void build_kv_self_defrag(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf) override; ggml_cgraph * gf) override;
// =======================================================
// === encoder-decoder === // === encoder-decoder ===
//
// TODO: this is temporary here, it will be moved
//
// whether we are computing encoder output or decoder output // whether we are computing encoder output or decoder output
bool is_encoding = false; bool is_encoding = false;
@ -497,23 +509,25 @@ protected:
struct ggml_tensor * inp_embd_enc; // F32 [n_embd, n_outputs_enc] struct ggml_tensor * inp_embd_enc; // F32 [n_embd, n_outputs_enc]
struct ggml_tensor * inp_kq_mask_cross; // F32 [n_outputs_enc, n_batch] struct ggml_tensor * inp_kq_mask_cross; // F32 [n_outputs_enc, n_batch]
virtual ggml_tensor * build_inp_embd_enc( ggml_tensor * build_inp_embd_enc(
ggml_context * ctx0) override; ggml_context * ctx0) override;
virtual ggml_tensor * build_inp_kq_mask_cross( ggml_tensor * build_inp_kq_mask_cross(
ggml_context * ctx0, ggml_context * ctx0,
int32_t n_tokens) override; int32_t n_tokens) override;
// ======================================================
// //
// state save/load // state save/load
// //
virtual size_t state_get_data(llama_io_write_i & io) override; size_t state_get_data(llama_io_write_i & io) override;
virtual size_t state_set_data(llama_io_read_i & io) override; size_t state_set_data(llama_io_read_i & io) override;
virtual size_t state_seq_get_data(llama_io_write_i & io, llama_seq_id seq_id) override; size_t state_seq_get_data(llama_io_write_i & io, llama_seq_id seq_id) override;
virtual size_t state_seq_set_data(llama_io_read_i & io, llama_seq_id seq_id) override; size_t state_seq_set_data(llama_io_read_i & io, llama_seq_id seq_id) override;
private:
// //
// members // members
// //
@ -532,24 +546,25 @@ public:
virtual ~llama_context_recurrent(); virtual ~llama_context_recurrent();
protected: protected:
virtual void reserve() override; void reserve() override;
public: public:
virtual llama_kv_cache * get_kv_self() override; llama_kv_cache * get_kv_self() override;
virtual const llama_kv_cache * get_kv_self() const override; const llama_kv_cache * get_kv_self() const override;
virtual void kv_self_update() override; void kv_self_update() override;
virtual int encode(llama_batch & inp_batch) override; int encode(llama_batch & inp_batch) override;
virtual int decode(llama_batch & inp_batch) override; int decode(llama_batch & inp_batch) override;
protected: protected:
// //
// input // input
// //
virtual void input_set(const llama_ubatch & ubatch) override; void input_set(const llama_ubatch & ubatch) override;
private:
struct { struct {
ggml_tensor * s_copy; // I32 [kv_size] ggml_tensor * s_copy; // I32 [kv_size]
ggml_tensor * s_mask; // F32 [1, n_kv] ggml_tensor * s_mask; // F32 [1, n_kv]
@ -559,20 +574,20 @@ protected:
// graph // graph
// //
virtual ggml_cgraph * graph_init() override; ggml_cgraph * graph_init() override;
public: public:
// //
// graph build // graph build
// //
virtual ggml_tensor * build_inp_s_copy( ggml_tensor * build_inp_s_copy(
ggml_context * ctx0) override; ggml_context * ctx0) override;
virtual ggml_tensor * build_inp_s_mask( ggml_tensor * build_inp_s_mask(
ggml_context * ctx0) override; ggml_context * ctx0) override;
virtual ggml_tensor * build_copy_mask_state( ggml_tensor * build_copy_mask_state(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * s, ggml_tensor * s,
@ -581,7 +596,7 @@ public:
int32_t n_state, int32_t n_state,
int32_t n_seqs) override; int32_t n_seqs) override;
virtual ggml_tensor * build_mamba_layer( ggml_tensor * build_mamba_layer(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * cur, ggml_tensor * cur,
@ -590,7 +605,7 @@ public:
const llama_ubatch & ubatch, const llama_ubatch & ubatch,
int il) override; int il) override;
virtual ggml_tensor * build_rwkv_token_shift_load( ggml_tensor * build_rwkv_token_shift_load(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * state_copy, ggml_tensor * state_copy,
@ -598,13 +613,13 @@ public:
const llama_ubatch & ubatch, const llama_ubatch & ubatch,
int il) override; int il) override;
virtual ggml_tensor * build_rwkv_token_shift_store( ggml_tensor * build_rwkv_token_shift_store(
ggml_context * ctx0, ggml_context * ctx0,
ggml_tensor * token_shift, ggml_tensor * token_shift,
const llama_ubatch & ubatch, const llama_ubatch & ubatch,
int il) override; int il) override;
virtual ggml_tensor * build_rwkv6_time_mix( ggml_tensor * build_rwkv6_time_mix(
ggml_context * ctx0, ggml_context * ctx0,
ggml_cgraph * gf, ggml_cgraph * gf,
ggml_tensor * cur, ggml_tensor * cur,
@ -619,12 +634,13 @@ protected:
// state save/load // state save/load
// //
virtual size_t state_get_data(llama_io_write_i & io) override; size_t state_get_data(llama_io_write_i & io) override;
virtual size_t state_set_data(llama_io_read_i & io) override; size_t state_set_data(llama_io_read_i & io) override;
virtual size_t state_seq_get_data(llama_io_write_i & io, llama_seq_id seq_id) override; size_t state_seq_get_data(llama_io_write_i & io, llama_seq_id seq_id) override;
virtual size_t state_seq_set_data(llama_io_read_i & io, llama_seq_id seq_id) override; size_t state_seq_set_data(llama_io_read_i & io, llama_seq_id seq_id) override;
private:
// //
// members // members
// //
@ -646,7 +662,7 @@ public:
virtual ~llama_context_enc_dec(); virtual ~llama_context_enc_dec();
protected: private:
llama_context_kv_self ctx_dec; llama_context_kv_self ctx_dec;
}; };

8
src/llama-graph.cpp
View File

@ -32,14 +32,6 @@ ggml_tensor * llama_graph_i::build_attn(
return nullptr; return nullptr;
} }
ggml_tensor * llama_graph_i::build_inp_self_k_shift(
ggml_context * ctx0) {
GGML_UNUSED(ctx0);
LLAMA_LOG_ERROR("%s: not implemented\n", __func__);
return nullptr;
}
ggml_tensor * llama_graph_i::build_inp_embd_enc( ggml_tensor * llama_graph_i::build_inp_embd_enc(
ggml_context * ctx0) { ggml_context * ctx0) {
GGML_UNUSED(ctx0); GGML_UNUSED(ctx0);

3
src/llama-graph.h
View File

@ -117,9 +117,6 @@ public:
float kq_scale, float kq_scale,
int il); int il);
virtual ggml_tensor * build_inp_self_k_shift(
ggml_context * ctx0);
virtual ggml_tensor * build_inp_embd_enc( virtual ggml_tensor * build_inp_embd_enc(
ggml_context * ctx0); ggml_context * ctx0);