ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

kv-cache : rework kv_cell (#13706)

Browse source 
* kv-cache : rework kv_cell

ggml-ci

* kv-cells : use "shift" instead of "delta" consistently

ggml-ci

* llama : add llama_max_parallel_sequences()

ggml-ci

* kv-cells : update comments [no ci]

* context : fail upon construction if sequences exceed max value

ggml-ci

* kv-cells : get_pos() -> pos_get() + comments

ggml-ci

* kv-cells : fix tracking of "used" cells

ggml-ci
This commit is contained in:
Georgi Gerganov 2025-05-25 16:34:36 +03:00 • committed by GitHub
parent c508256db2
commit de2ef53a4b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
8 changed files with 470 additions and 253 deletions
Download patch file Download diff file Expand all files Collapse all files

6
src/llama-context.cpp
View file

@ -25,7 +25,11 @@ llama_context::llama_context(
const auto & hparams = model.hparams;
cparams.n_seq_max = std::max(1u, params.n_seq_max);
cparams.n_seq_max = std::max(1u, params.n_seq_max);
if (cparams.n_seq_max > LLAMA_MAX_PARALLEL_SEQUENCES) {
throw std::runtime_error("n_seq_max must be <= " + std::to_string(LLAMA_MAX_PARALLEL_SEQUENCES));
}
cparams.n_threads = params.n_threads;
cparams.n_threads_batch = params.n_threads_batch;
cparams.yarn_ext_factor = params.yarn_ext_factor;