model: support MiMo-V2-Flash #18328
mimov2: convert ok
eb2cee1b
rename mimov2 --> mimo2
bd806505
fix conversion
86e8d1f1
runnable not incorrect
85937f90
use sink
4e810162
add_sliding_window_pattern
09d3df9b
add swa and per-layer n_head_kv
db8afa6a
correct params
237dfadf
somewhat working
55584783
correct gating func
3bf8f23c
nits
d4a3c4d4
ngxson
marked this pull request as ready for review 23 days ago
mimo2: wire RMS eps + MoE bias + converter guards
a5c54951
Merge pull request #69 from Aaryan-Kapoor/pr-18328
0f24c3b8
add co-author
e4761261
Merge branch 'master' into xsn/xiaomi_mimo_v2
d6f45334
CISC
approved these changes
on 2025-12-24
use add_rope_freq_base_swa
0cd227fe
ngxson
merged
4cbafad4
into master 22 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub