vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-05-11 04:51:25 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.github/workflows/build-linux-cross.yml
#	.github/workflows/build.yml
#	CODEOWNERS
#	ggml/CMakeLists.txt
#	ggml/src/ggml-cuda/fattn.cu
#	ggml/src/ggml-webgpu/CMakeLists.txt
#	ggml/src/ggml-webgpu/ggml-webgpu.cpp
#	ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl
#	tests/test-backend-ops.cpp
#	tests/test-chat-template.cpp
#	tools/llama-bench/llama-bench.cpp
#	tools/rpc/README.md
#	tools/server/README.md
This commit is contained in:
Concedo 2025-10-09 01:33:27 +08:00
commit b6f6338bba
32 changed files with 1556 additions and 636 deletions
Download patch file Download diff file Expand all files Collapse all files

9
src/llama-memory-recurrent.cpp
View file

@ -382,7 +382,9 @@ llama_memory_context_ptr llama_memory_recurrent::init_batch(llama_batch_allocr &
// if all tokens are output, split by sequence
ubatch = balloc.split_seq(n_ubatch);
} else {
ubatch = balloc.split_equal(n_ubatch, false);
// TODO: non-sequential equal split can be done if using unified KV cache
// for simplicity, we always use sequential equal split for now
ubatch = balloc.split_equal(n_ubatch, true);
}
if (ubatch.n_tokens == 0) {
@ -859,9 +861,12 @@ void llama_memory_recurrent::state_write_data(llama_io_write_i & io, const std::
bool llama_memory_recurrent::state_read_meta(llama_io_read_i & io, uint32_t cell_count, llama_seq_id dest_seq_id) {
if (dest_seq_id != -1) {
// single sequence
seq_rm(dest_seq_id, -1, -1);
if (cell_count == 0) {
return true;
}
llama_batch_allocr balloc(hparams.n_pos_per_embd());
llama_ubatch ubatch = balloc.ubatch_reserve(cell_count, 1);