Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

tests : add support for qwen3 SSM archs (#24031)

Browse Source 
* tests : add support for qwen3 SSM archs

* arch : add LLM_KV_ATTENTION_RECURRENT_LAYERS

* cont : naming + TODOs
This commit is contained in:
Georgi Gerganov
2026-06-03 10:15:27 +03:00
committed by GitHub
parent d545a2a993
commit 06938ac129
25 changed files with 109 additions and 83 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models/mellum.cpp
+1 -1
View File
@@ -13,7 +13,7 @@ void llama_model_mellum::load_arch_hparams(llama_model_loader & ml) {
if (res) {
hparams.set_swa_pattern(swa_period);
} else {
ml.get_key_or_arr(LLM_KV_ATTENTION_SLIDING_WINDOW_PATTERN, hparams.swa_layers, hparams.n_layer);
ml.get_key_or_arr(LLM_KV_ATTENTION_SLIDING_WINDOW_PATTERN, hparams.is_swa_impl, hparams.n_layer);
}
hparams.rope_freq_base_train_swa = hparams.rope_freq_base_train;