Superminaren/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-27 21:46:57 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into HEAD

Browse Source
This commit is contained in:
Georgi Gerganov
2025-11-29 22:38:44 +02:00
parent 9028ebfea8 8c32d9d96d
commit d8d98bb4bb
98 changed files with 6675 additions and 1609 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-context.cpp
+6 -2
View File
@@ -1,5 +1,6 @@
#include "llama-context.h"
#include "llama-arch.h"
#include "llama-impl.h"
#include "llama-batch.h"
#include "llama-io.h"
@@ -322,7 +323,7 @@ llama_context::llama_context(
cross.v_embd.clear();
const uint32_t n_seqs = cparams.kv_unified ? 1 : cparams.n_seq_max;
const uint32_t n_seqs = cparams.n_seq_max;
const uint32_t n_tokens = std::min(cparams.n_ctx, cparams.n_ubatch);
// avoid reserving graphs with zero outputs - assume one output per sequence
@@ -575,7 +576,7 @@ bool llama_context::memory_update(bool optimize) {
throw std::runtime_error("failed to initialize memory context");
}
const uint32_t n_seqs = cparams.kv_unified ? 1 : cparams.n_seq_max;
const uint32_t n_seqs = cparams.n_seq_max;
const uint32_t n_tokens = std::min(cparams.n_ctx, cparams.n_ubatch);
auto * gf = graph_reserve(n_tokens, n_seqs, n_tokens, mctx.get());
@@ -1849,6 +1850,9 @@ void llama_context::output_reorder() {
//
uint32_t llama_context::graph_max_nodes() const {
if (model.arch == LLM_ARCH_QWEN3NEXT) {
return std::max<uint32_t>(8192u, 32u*model.n_tensors());
}
return std::max<uint32_t>(1024u, 8u*model.n_tensors());
}