]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
model: support MiMo-V2-Flash (#18328)
authorXuan-Son Nguyen <redacted>
Wed, 24 Dec 2025 22:07:08 +0000 (23:07 +0100)
committerGitHub <redacted>
Wed, 24 Dec 2025 22:07:08 +0000 (23:07 +0100)
commit4cbafad4f09c82f1b32c76b714302f3937be1df7
treeed3e1bf72992069f85204050bd09f74c02c6bd8b
parentc18428423018ed214c004e6ecaedb0cbdda06805
model: support MiMo-V2-Flash (#18328)

* mimov2: convert ok

* rename mimov2 --> mimo2

* fix conversion

* runnable not incorrect

* use sink

* add_sliding_window_pattern

* add swa and per-layer n_head_kv

* correct params

* somewhat working

* correct gating func

* nits

* mimo2: wire RMS eps + MoE bias + converter guards

* add co-author

Co-authored-by: Aaryan-Kapoor <redacted>
* use add_rope_freq_base_swa

---------

Co-authored-by: Aaryan Kapoor <redacted>
Co-authored-by: Aaryan-Kapoor <redacted>
convert_hf_to_gguf.py
gguf-py/gguf/constants.py
gguf-py/gguf/tensor_mapping.py
src/CMakeLists.txt
src/llama-arch.cpp
src/llama-arch.h
src/llama-hparams.h
src/llama-model.cpp
src/llama-model.h
src/models/mimo2-iswa.cpp [new file with mode: 0644]
src/models/models.h