llama.cpp
add GLM-5.3-Flash (GLM5-Next) support
#27773
Merged

add GLM-5.3-Flash (GLM5-Next) support #27773

am17an merged 61 commits into ggml-org:master from timkhronos:GLM5.3-Flash
timkhronos
github-actions github-actions added model
github-actions github-actions added mtmd
github-actions github-actions added conversion
ngxson
ngxson commented on 2026-08-27
ngxson
ngxson commented on 2026-08-27
ngxson
ngxson commented on 2026-08-27
timkhronos Rebase GLM-Next support onto master, and migrate to llama-memory-hybr…
9370c82d
timkhronos timkhronos force pushed to 9370c82d 54 days ago
timkhronos Add initial MTP support
d4ce8259
danielhanchen
danielhanchen
Sciguy429
danielhanchen
timkhronos Merge branch optimizations. Reduce allocated compute buffer size, spe…
d2fc716c
timkhronos Review driven changes, remove env vars, protect tensors
98557115
timkhronos Strip MTP for initial PR
b93ed51f
timkhronos Clean up after mtp strip
ba436fd3
timkhronos Clean up after mtp strip
3ff8d92c
timkhronos Update speculative.cpp
f176c8a3
timkhronos Update llama-context.h
c34db581
timkhronos Clean up after mtp strip
0aa327b3
danielhanchen
timkhronos
danielhanchen
timkhronos Fix tokenizer ignore merges
ec1cdbf0
timkhronos Improve quantization protection selection
7152e9bf
timkhronos timkhronos marked this pull request as ready for review 52 days ago
timkhronos timkhronos requested a review 52 days ago
timkhronos timkhronos requested a review from CISC CISC 52 days ago
timkhronos timkhronos requested a review from ggerganov ggerganov 52 days ago
ngxson
ngxson commented on 2026-08-30
timkhronos Refactor mhc helpers, graph base
fdc54bed
timkhronos Merge branch 'master' into GLM5.3-Flash
85439412
ngxson
ngxson approved these changes on 2026-08-31
ngxson
timkhronos Lint Fixes
74bb0e39
timkhronos
CISC
CISC commented on 2026-08-31
timkhronos Apply suggestions from code review
a771613a
timkhronos Skip glm5-next in model saver, fix CRLF
5728a4bf
timkhronos Skip glm5-next in sweep
c35bddd0
timkhronos timkhronos requested a review from JohannesGaessler JohannesGaessler 50 days ago
github-actions github-actions added testing
timkhronos Remove T4 fallback
3bdb2d84
CISC
CISC commented on 2026-09-01
CISC
timkhronos Review cleanup
81f95af4
CISC
CISC commented on 2026-09-01
timkhronos Merge branch 'master' into GLM5.3-Flash
3498cc75
timkhronos Merge branch 'ggml-org:master' into GLM5.3-Flash
611e4070
timkhronos Review suggestions
1eca274e
CISC
CISC approved these changes on 2026-09-01
timkhronos Defer separate MTP gguf handling to MTP PR, drop filter
7de5a8e3
CISC
ggerganov
ggerganov commented on 2026-09-01
timkhronos Repad n_head_kv
9fe9fd71
timkhronos kpool init apply
a386cd76
timkhronos Merge branch 'ggml-org:master' into GLM5.3-Flash
5b05fccb
timkhronos Order by descending score
120eb9e4
CISC
CISC commented on 2026-09-02
timkhronos Drop guard
d0c55891
ggerganov
ggerganov commented on 2026-09-02
timkhronos read kpool from hparams, clarify kpool cache flags, remove kpool_buil…
8c289399
timkhronos Add glm5-next support to model saver and add arch test fixture
a2f1d20c
ggerganov
ggerganov commented on 2026-09-02
ggerganov
ggerganov commented on 2026-09-02
ggerganov
ggerganov commented on 2026-09-02
ggerganov
ggerganov commented on 2026-09-02
timkhronos Review cleanup
fe3187de
miminashi
timkhronos Kpool pooled caching clarify
2b533e09
Neresco
CISC
timkhronos Add multi stream support
ff6be954
timkhronos Merge branch 'ggml-org:master' into GLM5.3-Flash
ad74ec35
timkhronos Finish Rebase
1b564d2d
timkhronos
timkhronos Sparse FA fir DSA prefill
99fdaab4
timkhronos Merge branch 'ggml-org:master' into GLM5.3-Flash
db54719f
timkhronos Const
5c4bd50b
nicholasshirley
timkhronos Merge branch 'master' into GLM5.3-Flash
cc05bfee
nicholasshirley
timkhronos Update llama-model.cpp to fix rebase error
8134115f
LifesLight
CISC CISC requested a review from ggerganov ggerganov 44 days ago
fairydreaming
fairydreaming commented on 2026-09-09
sszymczy Merge remote-tracking branch 'upstream/master' into GLM5.3-Flash
05f7c98a
sszymczy gguf-py : merge tensor map entries for HC tensors
ae9893ff
sszymczy model : use build_gdn_l2_norm in GLM5_NEXT implementation
3f238d44
sszymczy chore : remove trailing whitespace
26ace38c
sszymczy model : use new OP precision setting API in GLM5_NEXT implementation
de253435
JPFlare
AIalliAI
am17an
am17an commented on 2026-09-12
am17an
am17an
am17an commented on 2026-09-12
pwilkin mtmd : use ggml_swiglu_clamp in GLM5V and apply the image token limit
df40ac60
pwilkin llama : keep the GLM5-Next k-pool layout across ubatches
166c62d5
pwilkin model : write the GLM5-Next recurrent rollback checkpoints
a22c8459
pwilkin Merge master into GLM5.3-Flash (PR #27773): resolve merge conflicts
36f338f1
pwilkin
pwilkin llama: fix PR #27773 test-save-load-state restore failure
93aefbcf
pwilkin llama: fix PR #27773 gpu-rocm graph reallocation
c2ed4857
am17an
pwilkin
pwilkin
am17an
pwilkin llama : fix GLM5-Next k-pool layout staleness after edits and shared …
1bceb45d
am17an
am17an approved these changes on 2026-09-28
ggerganov
ggerganov ggerganov assigned ggerganov ggerganov 22 days ago
pwilkin llama : fix build_attn_mha stream stride for non-contiguous q
18217f60
pwilkin llama : re-derive GLM5-Next k-pool sharing on state_read/state_drop
8583e617
pwilkin quant : drop duplicate GLM5-Next hc_ filter
2ad2809d
pwilkin Merge branch 'master' into GLM5.3-Flash (PR #27773)
544765ce
pwilkin
ggerganov
ggerganov commented on 2026-09-29
pwilkin glm5-next: scope K-pool cache access to indexed operations
4cab1fbf
pwilkin
ggerganov
ggerganov commented on 2026-09-29
pwilkin glm5-next: keep K-pool access in hybrid index memory
3746f394
ggerganov
pwilkin glm5-next: keep mHC graph builders model-local
8aaabc0b
pwilkin glm5-next: mark only touched pools per ubatch
1665c0e6
pwilkin
ggerganov
ggerganov approved these changes on 2026-09-30
am17an
ggerganov
am17an am17an merged 649dcb10 into master 20 days ago
ServeurpersoCom
ggerganov
adcape
alectrocute
ggerganov ggerganov added highlight
ichivangogh
am17an
ichivangogh
idsh22

Login to write a write a comment.

Login via GitHub

Assignees
Labels
Milestone