happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

cont : naming

This commit is contained in:
Georgi Gerganov 2025-11-30 00:07:13 +02:00
parent c187003d81
commit 80742cbaeb
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
6 changed files with 32 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
common/llguidance.cpp
View File

@ -112,10 +112,10 @@ static llama_sampler_i llama_sampler_llg_i = {
/* .reset = */ llama_sampler_llg_reset, /* .reset = */ llama_sampler_llg_reset,
/* .clone = */ llama_sampler_llg_clone, /* .clone = */ llama_sampler_llg_clone,
/* .free = */ llama_sampler_llg_free, /* .free = */ llama_sampler_llg_free,
/* .apply_ggml = */ NULL, /* .backend_init = */ NULL,
/* .accept_ggml = */ NULL, /* .backend_accept = */ NULL,
/* .set_input_ggml = */ NULL, /* .backend_apply = */ NULL,
/* .set_backend_context = */ NULL, /* .backend_set_input = */ NULL,
}; };
static size_t llama_sampler_llg_tokenize_fn(const void * user_data, const uint8_t * bytes, size_t bytes_len, static size_t llama_sampler_llg_tokenize_fn(const void * user_data, const uint8_t * bytes, size_t bytes_len,

4
include/llama.h
View File

@ -1374,10 +1374,6 @@ extern "C" {
// //
LLAMA_API struct llama_sampler * llama_sampler_init_infill(const struct llama_vocab * vocab); LLAMA_API struct llama_sampler * llama_sampler_init_infill(const struct llama_vocab * vocab);
//
// Backend samplers
//
// Returns the seed used by the sampler if applicable, LLAMA_DEFAULT_SEED otherwise // Returns the seed used by the sampler if applicable, LLAMA_DEFAULT_SEED otherwise
LLAMA_API uint32_t llama_sampler_get_seed(const struct llama_sampler * smpl); LLAMA_API uint32_t llama_sampler_get_seed(const struct llama_sampler * smpl);

6
src/llama-batch.cpp
View File

@ -29,7 +29,7 @@ bool llama_batch_allocr::init(
uint32_t n_embd, uint32_t n_embd,
uint32_t n_seq_max, uint32_t n_seq_max,
bool output_all, bool output_all,
bool backend_sampling) { bool sampling) {
clear(); clear();
batch = batch_inp; batch = batch_inp;
@ -146,7 +146,7 @@ bool llama_batch_allocr::init(
} }
} }
if (backend_sampling) { if (sampling) {
std::vector<int32_t> seq_output_count(n_seq_max, 0); std::vector<int32_t> seq_output_count(n_seq_max, 0);
for (int32_t i = 0; i < batch.n_tokens; ++i) { for (int32_t i = 0; i < batch.n_tokens; ++i) {
@ -157,7 +157,7 @@ bool llama_batch_allocr::init(
const llama_seq_id seq_id = batch.seq_id[i][s]; const llama_seq_id seq_id = batch.seq_id[i][s];
seq_output_count[seq_id]++; seq_output_count[seq_id]++;
if (seq_output_count[seq_id] > 1) { if (seq_output_count[seq_id] > 1) {
LLAMA_LOG_ERROR("%s: backend sampling allows at most one output token per sequence (%d)\n", __func__, seq_id); LLAMA_LOG_ERROR("%s: backend sampling requires at most one output token per sequence (%d)\n", __func__, seq_id);
return false; return false;
} }
} }

2
src/llama-batch.h
View File

@ -80,7 +80,7 @@ public:
uint32_t n_embd, uint32_t n_embd,
uint32_t n_seq_max, uint32_t n_seq_max,
bool output_all, bool output_all,
bool backend_sampling = false); bool sampling = false);
const llama_batch & get_batch() const; const llama_batch & get_batch() const;

23
src/llama-context.cpp
View File

@ -1322,12 +1322,12 @@ int llama_context::decode(const llama_batch & batch_inp) {
// when computing embeddings, all tokens are output // when computing embeddings, all tokens are output
const bool output_all = cparams.embeddings; const bool output_all = cparams.embeddings;
const bool has_backend_samplers = !sampling.samplers.empty(); const bool has_samplers = !sampling.samplers.empty();
if (!balloc->init(batch_inp, vocab, memory.get(), n_embd, if (!balloc->init(batch_inp, vocab, memory.get(), n_embd,
cparams.kv_unified ? LLAMA_MAX_SEQ : cparams.n_seq_max, cparams.kv_unified ? LLAMA_MAX_SEQ : cparams.n_seq_max,
output_all, output_all,
has_backend_samplers)) { has_samplers)) {
LLAMA_LOG_ERROR("%s: failed to initialize batch\n", __func__); LLAMA_LOG_ERROR("%s: failed to initialize batch\n", __func__);
return -1; return -1;
} }
@ -1415,10 +1415,6 @@ int llama_context::decode(const llama_batch & batch_inp) {
int64_t n_outputs_prev = 0; int64_t n_outputs_prev = 0;
// This flag indicates whether a backend sampler has actually sampled a specific
// token, or if it has produced probabilites. If true, we can skip the normal copying of logits and embeddings.
bool backend_has_sampled = false;
do { do {
const auto & ubatch = mctx->get_ubatch(); const auto & ubatch = mctx->get_ubatch();
@ -1477,9 +1473,11 @@ int llama_context::decode(const llama_batch & batch_inp) {
// ggml_graph_dump_dot(gf, NULL, "llama.dot"); // ggml_graph_dump_dot(gf, NULL, "llama.dot");
//} //}
backend_has_sampled = !res->t_sampled.empty() || !res->t_sampled_probs.empty() || !res->t_sampled_logits.empty(); // This flag indicates whether a backend sampler has actually sampled a specific
// token, or if it has produced probabilites. If true, we can skip the normal copying of logits and embeddings.
const bool has_sampled = !res->t_sampled.empty() || !res->t_sampled_probs.empty() || !res->t_sampled_logits.empty();
if (has_backend_samplers && backend_has_sampled) { if (has_samplers && has_sampled) {
const auto seq_to_output_row = build_seq_to_output_row(ubatch, n_outputs_prev); const auto seq_to_output_row = build_seq_to_output_row(ubatch, n_outputs_prev);
const auto stride = n_vocab; const auto stride = n_vocab;
@ -1495,7 +1493,6 @@ int llama_context::decode(const llama_batch & batch_inp) {
// async copy the candidate token ids from the backend to the host. // async copy the candidate token ids from the backend to the host.
// These are needed by CPU samplers to map probability/logit indices to vocab token ids. // These are needed by CPU samplers to map probability/logit indices to vocab token ids.
copy_tensor_async_candidates(res->t_candidates, sampling.candidates, stride, sampling.candidates_count, seq_to_output_row, sched.get()); copy_tensor_async_candidates(res->t_candidates, sampling.candidates, stride, sampling.candidates_count, seq_to_output_row, sched.get());
} }
auto * t_logits = res->get_logits(); auto * t_logits = res->get_logits();
@ -1661,7 +1658,7 @@ uint32_t llama_context::output_reserve(int32_t n_outputs, const llama_batch & ba
} }
// Check which sampling modes are needed by sequences in the current batch. // Check which sampling modes are needed by sequences in the current batch.
bool batch_has_backend_sampling = false; bool batch_has_sampling = false;
bool batch_needs_cpu_logits = false; bool batch_needs_cpu_logits = false;
if (batch.logits) { if (batch.logits) {
@ -1672,7 +1669,7 @@ uint32_t llama_context::output_reserve(int32_t n_outputs, const llama_batch & ba
for (int32_t j = 0; j < batch.n_seq_id[i]; j++) { for (int32_t j = 0; j < batch.n_seq_id[i]; j++) {
llama_seq_id seq_id = batch.seq_id[i][j]; llama_seq_id seq_id = batch.seq_id[i][j];
if (sampling.samplers.find(seq_id) != sampling.samplers.end()) { if (sampling.samplers.find(seq_id) != sampling.samplers.end()) {
batch_has_backend_sampling = true; batch_has_sampling = true;
} else { } else {
batch_needs_cpu_logits = true; batch_needs_cpu_logits = true;
} }
@ -1691,7 +1688,7 @@ uint32_t llama_context::output_reserve(int32_t n_outputs, const llama_batch & ba
logits_size = (has_logits && batch_needs_cpu_logits) ? n_vocab*n_outputs_max : 0; logits_size = (has_logits && batch_needs_cpu_logits) ? n_vocab*n_outputs_max : 0;
embd_size = has_embd ? n_embd*n_outputs_max : 0; embd_size = has_embd ? n_embd*n_outputs_max : 0;
if (!batch_has_backend_sampling) { if (!batch_has_sampling) {
sampling.logits_size = 0; sampling.logits_size = 0;
sampling.probs_size = 0; sampling.probs_size = 0;
sampling.sampled_size = 0; sampling.sampled_size = 0;
@ -1762,7 +1759,7 @@ uint32_t llama_context::output_reserve(int32_t n_outputs, const llama_batch & ba
embd = has_embd ? (float *) (base + offset) : nullptr; embd = has_embd ? (float *) (base + offset) : nullptr;
offset += embd_size * sizeof(float); offset += embd_size * sizeof(float);
if (batch_has_backend_sampling) { if (batch_has_sampling) {
sampling.logits = (float *) (base + offset); sampling.logits = (float *) (base + offset);
offset += sampling.logits_size * sizeof(float); offset += sampling.logits_size * sizeof(float);