tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-18 05:56:00 -04:00
Code Issues Packages Projects Releases Wiki Activity

llama : use n_swa + n_ubatch cells for SWA cache (#13833)

Browse Source 
* llama : use n_swa + n_ubatch cells for SWA cache

ggml-ci

* llama : add warning about multi-sqeuence SWA contexts
This commit is contained in:
Georgi Gerganov
2025-05-31 15:57:44 +03:00
committed by GitHub
parent c7e0a2054b
commit 3600cc2886
6 changed files with 24 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-kv-cache.cpp
View File

@@ -1731,14 +1731,14 @@ llama_kv_cache_unified_iswa::llama_kv_cache_unified_iswa(
bool swa_full,
uint32_t kv_size,
uint32_t n_seq_max,
uint32_t n_batch,
uint32_t n_ubatch,
uint32_t n_pad) : hparams(model.hparams) {
llama_kv_cache_unified::layer_filter_cb filter_base = [&](int32_t il) { return !model.hparams.is_swa(il); };
llama_kv_cache_unified::layer_filter_cb filter_swa = [&](int32_t il) { return model.hparams.is_swa(il); };
const uint32_t size_base = kv_size;
uint32_t size_swa = std::min(size_base, GGML_PAD(hparams.n_swa*n_seq_max + n_batch, n_pad));
uint32_t size_swa = std::min(size_base, GGML_PAD(hparams.n_swa*n_seq_max + n_ubatch, n_pad));
// when using full-size SWA cache, we set the SWA cache size to be equal to the base cache size
if (swa_full) {