Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-10 14:59:11 +02:00
Code Issues Packages Projects Releases Wiki Activity

mamba2: remove hardcoded 2x expansion factor and invalid d_inner % d_state check (#23082)

Browse Source 
* mamba2: remove hardcoded 2x expansion factor, support any expand value

* mamba2: remove invalid d_inner %% d_state check (unrelated parameters)

* Update convert_hf_to_gguf.py: make expand optional with default 2

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* mamba2: apply expand fix to refactored conversion/mamba.py

* also check for mamba_expand

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
Co-authored-by: Sigbjørn Skjæret <1629204+CISC@users.noreply.github.com>
This commit is contained in:
Arsen Arutunan
2026-06-26 08:50:54 +03:00
committed by GitHub
parent 5c7c22c3e1
commit 960d628f46
3 changed files with 10 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models/mamba-base.cpp
-1
View File
@@ -169,7 +169,6 @@ ggml_tensor * llm_build_mamba_base::build_mamba2_layer(llm_graph_input_rs * inp,
GGML_ASSERT(ubatch.equal_seqs());
GGML_ASSERT(ubatch.n_tokens == n_seq_tokens * n_seqs);
GGML_ASSERT(d_inner % n_head == 0);
GGML_ASSERT(d_inner % d_state == 0);
GGML_ASSERT(d_inner % n_group == 0);
ggml_tensor * conv_states_all = mctx_cur->get_r_l(il);