public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-09 15:55:37 +02:00
Code Issues Packages Projects Releases Wiki Activity

talk-llama : sync llama.cpp

Browse Source
This commit is contained in:
Georgi Gerganov
2026-02-08 09:29:10 +02:00
parent b0e81c1a2e
commit 4b23ff249e
26 changed files with 1661 additions and 179 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/talk-llama/llama-context.cpp
+3 -6
View File
@@ -317,6 +317,7 @@ llama_context::llama_context(
auto dev_type = ggml_backend_dev_type(ggml_backend_get_device(backend.get()));
if (dev_type == GGML_BACKEND_DEVICE_TYPE_CPU) {
// ignore CPU backend
// TODO: should we ignore ACCEL types too?
continue;
}
auto * dev = ggml_backend_get_device(backend.get());
@@ -1026,11 +1027,7 @@ bool llama_context::set_sampler(llama_seq_id seq_id, llama_sampler * sampler) {
llama_sampler_chain_n(sampler) > 0;
if (sampler && can_offload) {
ggml_backend_buffer_type_t buft = ggml_backend_dev_buffer_type(model.dev_output());
auto * host_buft = ggml_backend_dev_host_buffer_type(model.dev_output());
if (host_buft) {
buft = host_buft;
}
auto * buft = ggml_backend_dev_buffer_type(model.dev_output());
sampler->iface->backend_init(sampler, buft);
@@ -2016,7 +2013,7 @@ void llama_context::output_reorder() {
//
uint32_t llama_context::graph_max_nodes(uint32_t n_tokens) const {
if (model.arch == LLM_ARCH_QWEN3NEXT) {
if (model.arch == LLM_ARCH_QWEN3NEXT || model.arch == LLM_ARCH_KIMI_LINEAR) {
return std::max<uint32_t>(n_tokens * 40, 32u * model.n_tensors());
}
uint32_t res = std::max<uint32_t>(1024u, 8u*model.n_tensors());