starred/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-04-25 18:25:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

Update vendor ggml code to a5bb8ba4 (#13832)

Browse Source 
Co-authored-by: Daniel Hiltgen <daniel@ollama.com>
Co-authored-by: Gabe Goodhart <ghart@us.ibm.com>
Co-authored-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
This commit is contained in:
Jeffrey Morgan
2026-02-02 17:31:59 -08:00
committed by GitHub
parent 8f4a008139
commit ef00199fb4
241 changed files with 21271 additions and 5074 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
llama/llama.cpp/src/llama-context.h vendored
View File

@@ -40,6 +40,14 @@ struct llama_context {
~llama_context();
// reserve a new backend scheduler (if needed)
// for example, when:
// - changing loras
// - changing samplers
// - changing attention type
// - etc.
void sched_reserve();
void synchronize();
const llama_model & get_model() const;
@@ -70,6 +78,18 @@ struct llama_context {
float * get_embeddings_ith(int32_t i);
float * get_embeddings_seq(llama_seq_id seq_id);
llama_token * get_sampled_tokens() const;
llama_token get_sampled_token_ith(int32_t idx);
float * get_sampled_logits_ith(int32_t idx);
size_t get_sampled_logits_count(int32_t idx);
float * get_sampled_probs_ith(int32_t idx);
size_t get_sampled_probs_count(int32_t idx);
const llama_token * get_sampled_candidates_ith(int32_t idx);
size_t get_sampled_candidates_count(int32_t idx);
void attach_threadpool(
ggml_threadpool_t threadpool,
ggml_threadpool_t threadpool_batch);
@@ -192,10 +212,13 @@ private:
// Make sure enough space is available for outputs.
// Returns max number of outputs for which space was reserved.
uint32_t output_reserve(int32_t n_outputs);
uint32_t output_reserve(int32_t n_outputs, const llama_batch & batch);
void output_reorder();
// map the output row index `i` to batch index
int64_t output_resolve_row(int32_t i) const;
//
// graph
//
@@ -213,6 +236,8 @@ public:
ggml_cgraph * graph_reserve(
uint32_t n_tokens, uint32_t n_seqs, uint32_t n_outputs, const llama_memory_context_i * mctx, bool split_only = false, size_t * sizes = nullptr);
bool set_sampler(llama_seq_id seq_id, llama_sampler * sampler);
private:
llm_graph_params graph_params(
llm_graph_result * res,
@@ -252,6 +277,31 @@ private:
size_t embd_size = 0; // capacity (of floats) for embeddings
float * embd = nullptr;
// TODO: simplify
struct sampling_info {
std::map<llama_seq_id, llama_sampler *> samplers;
float * logits = nullptr;
size_t logits_size = 0;
llama_token * sampled = nullptr;
size_t sampled_size = 0;
float * probs = nullptr;
size_t probs_size = 0;
llama_token * candidates = nullptr;
size_t candidates_size = 0;
std::vector<uint32_t> logits_count;
std::vector<uint32_t> probs_count;
std::vector<uint32_t> candidates_count;
std::vector<llama_token> token_ids_full_vocab;
};
sampling_info sampling;
// sequence embeddings output (map of [n_embd] vectors)
// populated only when pooling_type != LLAMA_POOLING_TYPE_NONE
std::map<llama_seq_id, std::vector<float>> embd_seq;
@@ -272,6 +322,8 @@ private:
ggml_backend_sched_ptr sched;
bool sched_need_reserve = true;
ggml_backend_t backend_cpu = nullptr;
std::vector<ggml_backend_ptr> backends;