Commit Graph
1281 Commits
Author SHA1 Message Date
Lumpiasty 175eb7ac30 llama: free backend compute scratch on cold; guard memory_breakdown
On device release with KV eviction, also call ggml_backend_free_scratch()
on each backend so a cold model drops its Vulkan compute preallocations
(reallocated lazily on the next compute via restore).

Also guard llama_context::memory_breakdown() against a freed scheduler:
evict_kv release resets sched to null, so a cold model that is then torn
down (e.g. terminated by the process manager) hit GGML_ASSERT(sched) in
ggml_backend_sched_get_buffer_type. Skip the compute-buffer accounting
when sched is null.

Assisted-by: Claude
2026-10-09 00:32:05 +02:00
Lumpiasty 4b553876cd llama: evict recurrent/SSM state on device release
The recurrent (SSM/conv) state of hybrid models (e.g. Qwen3.5) was left
resident when a model's device buffers were released for on-demand VRAM
sharing - llama_memory_recurrent::release_device_buffers() was a no-op
default. Implement it (and restore_device_buffers) with the same
capture-host-shadow / free / reallocate pattern as llama_kv_cache, so
llama_memory_hybrid now evicts both its attention KV and its recurrent
state. The state is read-write, so its shadow is recaptured on every
release.

Assisted-by: Claude
2026-10-09 00:31:39 +02:00
Lumpiasty c68f58f251 server: also free the compute-graph scheduler on KV eviction
Under LLAMA_SLEEP_EVICT_KV a cold model still held the scheduler's worst-case
compute buffer (hundreds of MiB, e.g. ~700 MiB for gemma-26B) plus the resident
experts. release_device(evict_kv) now also frees the sched (sched.reset() +
sched_need_reserve), and restore_device() rebuilds it via sched_reserve(), so a
fully-evicted model holds essentially no VRAM (gemma cold: 7147 -> 51 MiB).
Trades a heavier re-warm (weights+KV H2D + sched reserve) for the freed VRAM;
weights-only mode is unchanged (fast switch, keeps KV+compute).

Assisted-by: Claude
2026-10-09 00:31:17 +02:00
Lumpiasty 9d266b5317 llama: guard buffer iteration against released device buffers
release_device_weights()/release_device_buffers() leave a null buffer in
ctxs_bufs while the device memory is released for on-demand VRAM sharing.
The memory-breakdown / total_size / clear paths iterated these and called
ggml_backend_buffer_get_size()/get_type()/clear() on the null buffer, tripping
GGML_ASSERT(buffer) and aborting on shutdown of a cold model. Skip null buffers
in llama_model::memory_breakdown and llama_kv_cache::{memory_breakdown,
total_size,clear}.

Assisted-by: Claude
2026-10-09 00:31:17 +02:00
Lumpiasty aeac8898d5 server: optionally evict the KV cache too (Phase 2 of VRAM sharing)
Extend on-demand device residency to the KV cache so that when a model's KV
plus another model would not fit in VRAM, the KV can also be evicted to a host
shadow (D2H on release, H2D on restore) instead of only the weights.

- llama_memory_i: add release_device_buffers()/restore_device_buffers()
  (default no-op). Implemented in llama_kv_cache (D2H shadow of the live
  ctxs_bufs, freed and reallocated like the weights); llama_memory_hybrid and
  llama_kv_cache_iswa delegate to their child caches.
- llama_context::release_device(evict_kv): also evict the memory's device
  buffers when requested; restore_device() rebuilds them. Public API
  llama_context_release_device gains an evict_kv flag.
- server: LLAMA_SLEEP_EVICT_KV=1 enables it. Off by default (weights-only),
  since the KV shadow adds a D2H/H2D copy of the live cache each cycle.

Validated on RX 580 (Vulkan), 4B @ 32k ctx: weights-only cold VRAM 1750 MB
(KV stays); weights+KV cold VRAM 726 MB (KV freed, ~1 GB reclaimed). KV
survives the round-trip: prompt cache reused after the cycle (prompt_n 4 vs
42), correct output.

Assisted-by: Claude
2026-10-09 00:31:17 +02:00
Lumpiasty 1e487057b8 server: on-demand VRAM sharing to time-share one GPU between models
Add release/restore of a model's GPU weight buffers (keeping a host shadow
and the KV cache) so several always-loaded llama-server processes can
time-share a single GPU without reloading or losing the prompt cache.

- llama-model: release_device_weights()/restore_device_weights() capture a
  compact host shadow (stable iteration order, view-skipping) and free then
  realloc the device weight buffers; weights_resident() query.
- llama-context: release_device()/restore_device() wrappers; decode() auto-
  restores; public C API llama_context_release_device/restore_device.
- server: LLAMA_SLEEP_VRAM_ONLY makes idle-sleep release only the VRAM weights
  (not a full unload/reload). A cross-process flock token in LLAMA_VRAM_ARENA
  enforces "resident iff holds token"; an inotify doorbell forces the holder
  to release on contention. The warden thread only touches the task queue, so
  releases run on the loop thread and never race a decode.

Validated on RX 580 (Vulkan): two models share 8GB, never both resident,
correct output under contention, KV cache preserved (no re-prefill).

Assisted-by: Claude
2026-10-09 00:31:07 +02:00
LumpiastyandClaude Opus 5.5 b77fb98acb llama: skip the routing-ids readback in sched_copy_experts when the batch uses every expert
Offloaded expert weights are uploaded per split, and to decide which experts to
upload the copy callback reads the routing ids back to the host. The ids are
produced on the same device we are about to upload to, so the readback forces a
full pipeline flush - 47 of them per eval on a 48-layer MoE.

Once the batch draws enough experts the readback stops telling us anything. At
2048 tokens x 10 experts over 256 experts every expert comes back used, so the
used set is all true and the copy is a single whole-tensor range anyway.

Skip the readback when ggml_nelements(ids) >= 4 * n_expert and mark every expert
used. Uploading an expert that no id selects cannot change the result, since
mul_mat_id only reads the rows the ids point at, so this stays exact. With the
MoE cache on, the all-used set still goes through moe_cache->copy_experts, which
copies the cached experts device-to-device from the host slot maps and uploads
the rest - still exact, and no readback is needed for it. Decode is unaffected:
n_ids there is the number of experts per token, far below the threshold. The
small-batch MoE cache slot path has its own readback and is not touched.

The st.ids/st.used caching is unchanged: a repeat call with the same ids tensor
reuses the all-true set, which is what the threshold would decide again.

Port of dbc449afd ("sched: skip the MoE routing-ids readback when the batch
uses every expert"), whose target block in ggml_backend_sched_compute_splits was
removed upstream (#29943) in favour of this copy callback.

Measured (original, before the port) on an RX 580 (Vulkan, Polaris) running
Laguna-S-2.1 118B IQ2_M with all experts host-resident: pp2048 +4.7% at depth
32k, +0.9% at 16k, tg32 +3%.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PZz44SLQvTXMyWGio6t9DZ
2026-10-09 00:30:07 +02:00
Hrishith Thadicherla a11f57ba93 model : fix DFlash output head sharing (#30111)
* llama : fix DFlash output head sharing

Assisted-by: Codex

* dflash : read tied output weights from GGUF metadata

Assisted-by: Codex

* llama : share tied word embedding metadata

Assisted-by: Codex

* llama : remove DFlash embedding head fallback

Assisted-by: Codex
2026-10-08 17:42:28 +02:00
Aman Gupta c811cb8f0a llama: support MoE cache over multiple GPUs (#30112) 2026-10-08 15:50:49 +05:30
bosh 37ac634566 model : support classifier_activation for rerankers (#29692)
* model : support classifier_activation for rerankers

Assisted-by: Claude Opus 5.5

* model : map classifier gelu to gelu_erf and accept tanh

Assisted-by: Claude Opus 5.5

* model : default act_cls to tanh, ModernBERT falls back to gelu_erf

Assisted-by: Claude Opus 5.5
2026-10-08 09:29:20 +03:00
Tarek Dakhran a657f7e981 model : add LiquidAI/d1-omni-600M decision model (#30114)
* model : add LiquidAI/d1-omni-600M decision model

Assisted-by: Claude Opus 5.5

* mtmd : keep conformer GLU sigmoid on CUDA

Assisted-by: Claude Opus 5.5

* server : take d1omni audio through images and input_audio, scope memory-less lfm2 to non-causal

Assisted-by: Claude Opus 5.5

* common : rename decision type d1omni to lfm2-d1-omni, server : make images an alias of files

Assisted-by: Claude Opus 5.5
2026-10-08 02:07:49 +02:00
Aman Gupta d6cf9acb25 llama : add a GPU cache for MoE experts kept in host memory (#29887)
* llama : add a GPU cache for MoE experts kept in host memory

Assisted-by: Claude

* use llama_moe_cache_ptr
2026-10-07 21:07:45 +03:00
Pascal 448147d42a llama: share the nextn tensor flags between models (#30097)
* llama: share the nextn tensor flags between models

Follow-up of the TODO in glm5-next: move the trunk-only and MTP-only
detection that each model copied into a nextn_flags helper of
llama_model_base. It probes the first trunk layer and the first NextN
layer, and adds TENSOR_SKIP when MTP is not loaded. qwen4exp probes
hc_attn_norm since it has no attn_norm.

deepseek4, nemotron-h, qwen35, qwen35moe, qwen3next and qwen4exp now
also accept a trunk-only file, like the other models.

* llama: avoid capturing structured bindings in the nextn flags

Lambdas that capture structured bindings need C++20, and GCC 15
rejects them under -Werror, so the models read the trunk and MTP
flags into plain variables.
2026-10-07 16:30:00 +02:00
b9acf138a1 feat: add GLM5Next MTP, optimize (#29928)
* llama : add GLM5-Next NextN (MTP) graph

Build the GLM5-Next multi-token-prediction head as graph_mtp: the NextN block
embeds enorm(tok)+hnorm(h) through eh_proj, runs one plain DSA layer and the
shared lm_head, reusing the trunk's builders through the no_build tag ctor.
llama_memory_recurrent also tolerates a partial seq_rm when the context holds
no recurrent layers, which is what the MTP draft context needs.

Assisted-by: Claude

* llama : glm5-next: skip dead compute in headless NextN forwards

A NextN forward with no output rows (the MTP catch-up and the draft-context
prefill) persists only through its cache writes, so the headless graph keeps
the MLA latent, indexer key|gate and pooled-key writes and drops the query
path, the indexer selection, the attention body, the FFN and the LM head. The
4-token catch-up falls from 6.9 ms to 0.33 ms of kernels; the greedy output
hashes and the draft acceptance are unchanged.

Assisted-by: Claude

* llama : glm5-next: fix NextN extraction contracts and shared-tail rollback

Three fixes from the architectural review. The headless graph prune now also
requires that no unmasked nextn extraction is live, because that mode reads
n_tokens hidden rows regardless of the logits flags. Masked extraction
publishes the hidden rows gathered by the output ids, so a batch whose output
flags are not a prefix exports the right rows. A partial recurrent rollback
whose tail cell is shared with another sequence is now rejected instead of
silently moving that sequence's tail.

Assisted-by: Claude

* llama : glm5-next: tidy comments in the MTP changes

Assisted-by: Claude

* llama : glm5-next: crop the MTP graph to the output rows instead of pruning it

Replace the headless NextN prune with the crop pattern the other MTP
graphs use: gather the attention output and the block input at the
output ids before the position-wise FFN and the shared head. A NextN
forward with no output rows (the MTP catch-up and the draft-context
prefill) then runs the FFN and the head over zero rows. The 4-token
catch-up falls from 6.9 ms to 2.9 ms of kernels; greedy output hashes
are unchanged.

Assisted-by: Claude

* glm5-next: use the nextn crop helpers in the MTP graph

Replace the local crop condition and the masked select of t_h_nextn
with crop_before_nextn and crop_after_nextn, so the MTP graph narrows
its rows the same way as the main graph and the other models.

Describe the shared cell and empty filter branches of the recurrent
partial rollback.

* glm5-next: load MTP-only and trunk-only GGUF files

Make the trunk tensors optional when the file only holds the NextN
layer, and the NextN tensors optional when the file only holds the
trunk, so the split MTP GGUF loads as a draft model.

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Pascal <admin@serveurperso.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-10-07 13:16:53 +02:00
Hrishith ThadicherlaandGeorgi Gerganov d0b490f25e sampling : use greedy selection for eligible temperature-zero chains (#29797)
* sampling : use greedy selection for eligible temperature-zero chains

Assisted-by: OpenAI Codex

* Apply suggestion from @ggerganov

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* sampling: simplify zero-temperature greedy eligibility

Allow the same greedy selection on CPU and grammar/reasoning-budget paths.
Keep distribution sampling for dynamic temperature and requested probabilities.
Cover the common sampler selection and probability behavior in the existing sampler tests.

Assisted-by: OpenAI Codex

* sampling : use greedy selection after final top-k with k=1

Assisted-by: OpenAI Codex

* cont : clean-up

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-10-07 13:34:18 +03:00
Sigbjørn Skjæret 42b021b4dc vocab : add plamo fim tokens (#30090) 2026-10-07 11:40:35 +02:00
Toki NasinandSigbjørn Skjæret abeada335e vocab : implement PLaMo-3 tokenizer pre-segmentation (#30045)
* vocab : implement PLaMo-3 tokenizer pre-segmentation

The PLaMo-3 tokenizer inserts hard boundaries before running the Unigram
DP, around <|plamo:...|>-looking text, and around runs of at least 4
identical characters or 2 spaces. Without them llama.cpp tokenizes code
indentation and repeated punctuation differently from the reference.

Reproduce the two re.sub() passes in llm_tokenizer_plamo2 by encoding
each segment independently.

* add vocab type "plamo3"

* Update src/llama-vocab.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

* misc change

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
2026-10-06 19:07:54 +02:00
4625240437 model : add K2 Horizon dense and MoVA support (#29535)
* model: K2 Horizon gguf conversion code

* model: loading hparams and tensors in k2-horizon.cpp

* model: K2 Horizon compute graph

* model: K2 Horizon compute graph adjustment and registering tokenizers

* model: K2 Horizon chat template and accomodate safetensors naming

* unicode : add the K2-Horizon pre-tokenizer splitter

The K2-Horizon regex had no arm in unicode_regex_split_custom and fell through to the
general std::regex fallback, which fails two ways.

On MSVC std::regex rejects \p{...}, so no K2-Horizon GGUF loads on Windows at all:
llama-quantize, llama-imatrix and llama-perplexity all abort with
regex_error(error_escape) before a token is produced.

Where the fallback does compile it is still wrong. unicode_regex_split collapses each
codepoint to a single byte naming its Unicode category before matching, and U+200C/U+200D
are category Control, which has no entry in k_ucat_cpt, so both become the 0xD0 fallback
byte. The literal ‌ and ‍ alternatives in K2's regex can then never match and
every ZWNJ or ZWJ ends a letter run.

The splitter is the existing llama3 one with a single rule widened, since K2's regex
differs from llama3's only in that a letter run also takes marks, ZWNJ and ZWJ.

tests/test-unicode.cpp gains a case for this: it fails before the change with
[Amy] [ZWNJ khaham] and passes after with the run intact.

* tests: expand K2 Horizon unicode splitter coverage

* unicode: handle K2 Horizon case folding and empty input

Assisted-by: Codex

* jinja : support sequence indices in selectattr and rejectattr

Assisted-by: Codex

* model : add K2 Horizon dense and MoVA support

Includes the K2 Horizon implementation from ifm-ai/llama.cpp with converter, tensor-parallel and model save/reload fixes.

Assisted-by: Codex

* chat : support K2 Horizon reasoning and tool calls

Assisted-by: Codex

* conversion: remove obsolete K2 Aurora alias

Assisted-by: Codex

* k2-horizon: enforce response schemas and load YaRN betas

Constrain final JSON after reasoning, accept flexible JSON tool envelopes,
enforce XML dialects, and handle repeated or alternate thinking markers.
Load YaRN beta metadata instead of retaining the default values.

Add schema, streaming, continuation, and model reload regressions. Validate
CUDA and CPU builds and 0.9B, 4B, and MoVA conversation/tool round trips.

Assisted-by: Codex

* renaming template fixture

* adressing cisc follows ups

* desloppify the parser / adress aldehir comments

* clean test-chat

* remove fallback : model trained mostly on high anyway

* fix k2 attn_v_exp tn splitting and metal fusion baseline

* k2-horizon : forward expand views before sums

* k2-horizon: copy embds before group norm to fix TP

* disable tesnor parallelism

---------

Co-authored-by: Ryandito Diandaru <ryandito.diandaru@mbzuai.ac.ae>
Co-authored-by: WestWaters <mario.papaleo2013@gmail.com>
Co-authored-by: Natani L. Mayday <71436458+TaskPuppyNatani@users.noreply.github.com>
Co-authored-by: West <100190545+WestWaters@users.noreply.github.com>
Co-authored-by: aaryamonvikram <aaryamonvikram@gmail.com>
Co-authored-by: aaryamonvikram <96529820+aaryamonvikram@users.noreply.github.com>
2026-10-06 18:42:24 +02:00
Pascal 3109914090 llama: remove the gather path of the glm5-next sparse attention (#30042)
The gather path attended over the selected latents with a plain
matmul and softmax. It only ran with n_ubatch <= 16, and the flash
attention backends now skip the masked rows through n_kv_max, so
the scatter path covers every case.

Drop the gather flag, the gathered attention branch and
gather_mla_rows. set_input_kpool always maps padding to the n_kv
sentinel, and the slot mask becomes sel_mask since only the
scatter reads it.
2026-10-06 18:31:13 +02:00
Xuan-Son Nguyen 4fbc76dec5 model: support embeddinggemma2 (text+vision+audio) (#30054) 2026-10-06 18:20:31 +02:00
Georgi Gerganov f0c41e0168 models : consolidate nextn row cropping into shared helpers (#30017)
* mimo2 : always emit h_nextn

the other nextn-capable models set it unconditionally

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD

* models : consolidate nextn row cropping into shared helpers

- replace the duplicated crop conditions and the per-model flags (narrow_early,
  crop_before_ffn, crop_last_layer, emit_h_nextn) with two helpers on llm_graph_context:
  crop_before_nextn() / crop_after_nextn()
- models that only tested embeddings_nextn_masked now share the same condition, so they
  crop the last layer before the nextn capture whenever extraction is off
- t_h_nextn is now set unconditionally in mimo2, qwen4exp and deepseek4 (as in the other
  nextn-capable models); host-side reads stay gated by cparams.embeddings_nextn

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD
2026-10-06 14:08:54 +03:00
Piotr Wilkin (ilintar) 1a3011cc0c llama : re-reserve the sched when the nextn extraction flags change (#30020)
The speculative MTP init enables NextN extraction on the target and draft
contexts after both were created and their schedulers reserved. With
unmasked extraction the trunk graph keeps every token through the last
layer instead of cropping to the output rows, so the first decode
reallocates to that batch's shape and the next, wider batch trips
GGML_SCHED_DEBUG_REALLOC. Invalidate the reserve when the flags change so
the next compute re-reserves with the new graph shape.

Assisted-by: Claude
2026-10-06 10:41:23 +02:00
Aman GuptaandGeorgi Gerganov 6753a033f0 ggml: refactor selective expert copying to user code (#29943)
* ggml: refactor selective expert copying to user code

* tests: enroll two models into selective expert copy test

* tests: use deepseek2 as test model

* improve comment in ggml-backend.h

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* cont: fix whitespace

* cont : better comments

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-10-06 10:59:45 +03:00
Pascal 43fe9c6428 llama: fix k-pool scatter data race on shared sequences (#29994)
* llama: re-pool each shared k-pool rep once

With shared cells every pool is re-pooled, and since the pooled keys
are always scattered, the pools a seq_cp shares between sequences
wrote the same rep row from several scatter entries, a data race on
the CPU backend. Mark each rep once: the sharing sequences read the
same row through pool_cells.

* llama: assert whole-sequence seq_cp in the hybrid idx memory

The recurrent state is always copied whole whatever the range, and a
k-pool cell shared by a partial copy could carry two pool groupings
with a single pooled row. Every caller copies whole sequences, so
reject partial ranges instead of supporting them.

* llama: drop the k-pool cache_safe mode

With whole-sequence seq_cp, sequences sharing cells share their pools
too, so the pooled row of a shared rep is valid for all of them. Mark
each rep once in every ubatch instead of re-pooling everything while
cells are shared, which removes the sharing scan and the stale-all
workarounds in seq_rm, state_read and state_drop. seq_cp now only
stales the destination.
2026-10-06 07:41:57 +02:00
Xuan-Son Nguyen 9871df5911 server: support vision input for Clef (#29969)
* server: support vision input for Clef

* move input_attn_causal to private

* extend old server_batch::embd

* server_batch::token::pos to multi dim

* nits

* fix abort

* fix img tokens cap

* fix yield_to_queue mutate data
2026-10-05 14:02:40 +02:00
Georgi GerganovandPascal c173a53bdf llama : fix unexpected graph reallocation in the k-pool models (#29958)
* llama : fix unexpected graph reallocation in the k-pool models

Both k-pool models built a graph shape that depends on state the
full-context reserve cannot know:

- qwen4exp branched on inp->cache_safe, which turns false as soon as
  llama_memory_seq_cp shares cells (e.g. batched-bench -pps): the QSA
  layers swapped scatter+gather for fill+concat and dropped the
  new_pool_rep leaf, so the decode graph had 12 fewer nodes than the
  reserved one
- glm5-next branched on gather = n_tokens <= 16 && n_kv > n_sel, so the
  TG decode built the gather shape (7564 nodes) while the last reserve,
  the PP one, had the dense shape (7762 nodes)

Either mismatch forces a decode-time re-reserve that drops the
worst-case sizing and bakes in the current state, so the next state
growth (n_pool, n_kv, n_new) needs more room at an unchanged graph size
and aborts under GGML_SCHED_DEBUG_REALLOC=1. Reproduce with, e.g.:

  GGML_SCHED_DEBUG_REALLOC=1 ./bin/llama-batched-bench \
    -hf ggml-org/GLM-5.3-Flash-GGUF:Q2_K -npp 2500 -ntg 32 -npl 1,2 \
    -c 32768 -pps -kvu

Always scatter+gather the pooled keys, and pick gather from context
constants only: n_ubatch bounds every ubatch, top_k + kpool - 1 bounds
n_sel. Every graph of a context then shares one shape, which the
reserve covers, and the dense path measured faster than the gather path
at 2.5k and 16k context.

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD

* llama : drop the unused k-pool cache_safe graph API

The k-pool graphs no longer branch on cache_safe, so nothing reads
get_kpool_cache_safe() or the conditional new_pool_rep any more: both
models always pass the scatter target, which set_input_kpool now
requires instead of merely preferring.

Also drop the cache_safe copy in kpool_build_sizes(), a sizes-only
helper. The layout and state flag itself stays, it still decides which
pools a layout with shared cells must re-pool.

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD

* tests : add a shared-seq graph reserve regression test

Decode a prompt into seq 0, share its cells with seq 1 via
llama_memory_seq_cp (what llama-batched-bench does for -pps), then keep
decoding both sequences. For the k-pool models sharing clears
cache_safe, which changes the graph topology while the pools keep
growing, so a scheduler that re-reserves with the current state
instead of the worst-case one aborts under GGML_SCHED_DEBUG_REALLOC=1.
The test registration sets that flag, and the test aborts on both
k-pool models before 2220411ec1.

kimi-linear and minimax-01 are skipped: they reserve the final pp graph
with n_seqs = 1 (see [TAG_RESERVE_DIAG_DECAY] in llama-context.cpp), so
every multi-seq graph has a different layout and re-reserves by design.

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD

* cont : add TODOs

* cont : fix comment

* cuda: match the moe weighted reduction on empty ubatches

ggml_cuda_match_moe_weighted_reduction rejected tensors with zero
rows. A ubatch without outputs shrinks the last layer to zero rows
through inp_out_ids, so graph_optimize dropped its alloc dep there and
the scheduler graph lost one node compared to the reserved one. The
scheduler then re-reserved at the size of that ubatch, and the next
ubatch with the same node count but larger tensors aborted under
GGML_SCHED_DEBUG_REALLOC=1.

The compute loop already skips empty nodes before trying any fusion,
so the guard only made the alloc deps depend on the row count.

* tests: build the rollback test only where internal symbols link

The shared-seq case calls llm_arch_from_string, which libllama does
not export through LLAMA_API, so linking test-recurrent-state-rollback
fails on Windows with shared libraries. Its build now sits in the
NOT WIN32 OR NOT BUILD_SHARED_LIBS block, next to test-llama-archs and
the test registration it already lives under.

* tests: skip archs by name in the shared-seq reserve test

The skip of kimi-linear and minimax-01 went through llm_arch_from_string,
which libllama does not export through LLAMA_API, so the test could not
link on Windows with shared libraries. It now compares the
general.architecture string directly, and the test builds on every
platform again.

---------

Co-authored-by: Pascal <admin@serveurperso.com>
2026-10-05 11:36:25 +03:00
Evan HuusandGeorgi Gerganov 210791069b kv-cache: fix restoring mismatched KV cache rotation by saving exact rotation metadata (#28498)
* kv-cache: save exact KV rotation metadata, reject restoring mismatched rotation

* tests : move the state rotation test to test-save-load-state

the test is now part of the save/load test matrix and runs against
every model under test, like the rest of the suite

it probes the KV cache type combinations supported by the model and
treats models that do not use attention rotation as passing vacuously

Assisted-by: pi:llama.cpp/Qwen3.8-27B

* cont : skip unsupported KV caches

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-10-05 11:27:30 +03:00
Xuan-Son Nguyen 0bb496dbd3 llama: support both embd + raw tokens in batch (#29622)
* llama: support both embd + raw tokens in batch

* add to test-llama-archs

* also check case llm_arch_supports_mixed_batch = false

* constant graph topology

* have dedicated input for mixed case

* rm set_tensor_backend

* is_embd --> type

* consolidate m-rope pos handling into one place

* nits
2026-10-05 01:35:49 +02:00
Pascal 436f6f89e1 graph: gather the recurrent states once so the reserve covers every split (#29856)
build_rs gathered the extra states (n_rs - n_seqs rows) with their own
get_rows. The worst-case reserve has n_rs == n_seqs, so that node was
sized at zero rows, and any ubatch whose cells are not contiguous forced
a graph reallocation at an unchanged node count, which aborts under
GGML_SCHED_NO_REALLOC.

A single get_rows now gathers the n_rs states: the ubatch states and the
extra states are views of it, and its size only depends on n_rs, which
the reserve already sets to the maximum. A custom getter (mamba ssm_scan)
gathers from the second state, so a single sequence ubatch copies no
state. The views are built once per graph in the input to keep the host
overhead of the graph unchanged.
2026-10-03 14:02:05 +02:00
PascalandRuben Ortlam 889edf43dd qwen4exp : halve the indexer score memory (#29825)
* qwen4exp : halve the indexer score memory

The indexer scored all heads in one product and rectified a copy of it,
so two [n_pool, n_idx_h, n_tokens] f32 tensors were live at once, the
largest buffers of the graph at long context. Each head now gets its
own product, rectified and summed in place into one [n_pool, n_tokens]
score.

* qwen4exp: let the allocator reuse the indexer score buffers

Address review from CISC: use plain ggml_add and ggml_relu in the
indexer head loop. The graph allocator already runs them in place when
their source has no other consumer, so the _inplace variants are not
needed. The compute buffer and the speed are unchanged.

* cuda: support 4 heads in the lightning indexer

Dispatch 4 heads to the vector kernel, too few for a wmma tile, and
accept them in supports_op. test-backend-ops covers 4 heads.

* metal: take the lightning indexer head count as a function constant

The kernel reads the head count from a function constant and zero fills
the last head tile, so any head count runs and 64 heads is unchanged.

* qwen4exp: compute the indexer score with the lightning indexer

Address review from am17an: the unweighted sum of the rectified head
scores scaled by 1/sqrt(head_dim) is the lightning indexer with every
head weight set to that scale, so the indexer calls
ggml_lightning_indexer on the pooled keys with an f16 pool mask. The
keys are read once for all heads and no per head score is
materialized.

* vulkan: tile the lightning indexer over keys and tokens

A workgroup scores 64 keys against 8 tokens: the keys are staged once
in shared memory, the queries one head at a time, and each invocation
owns one key for two tokens, so no dot product needs a cross invocation
reduction. The subgroup variant and the flat dispatch are gone, the grid
is keys x tokens x streams.

* vectorize vulkan loads and use fp16 dot product

---------

Co-authored-by: Ruben Ortlam <rortlam@redhat.com>
2026-10-03 07:19:00 +02:00
Xuan-Son NguyenandSigbjørn Skjæret 99b95488ca model: add support for clef decision model (text-only) (#29831)
* init support for clef (text only)

* more static graph

* clean up

* nits

* nits 2

* Update gguf-py/gguf/constants.py

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
2026-10-03 02:50:48 +02:00
Xuan-Son Nguyen a4cb4c61fd llama, server: add /v1/systemone API (models: laya, julia-1, lev, openjev, kev) (#29818)
* init conversion

* convert: ok

* model loaded

* add server code

* improve conversion script

* support shared prompt prefix

* add docs, imorove UX a bit

* add vision support

* add openjev tiny model for testing

* add dev docs

* support lev & kev

* clean up

* fix lev noul

* fix py lint

* nits docs

* clarify about not supporting date_facts
2026-10-02 11:56:04 +02:00
Adrien Gallouët 8d81559fa7 llama : silence unused-result warnings (#29839)
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-10-02 11:07:06 +02:00
Adrien Gallouët 6805ae35df llama : use GGML_ABORT instead of throw (#29840)
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-10-02 10:44:44 +02:00
Georgi Gerganov 4e2713c162 qwen4exp : optimize mask constructions (#29824)
* qwen4exp : optimize mask constructions

* cont : apply the same change for GLM5-next
2026-10-02 11:08:50 +03:00
Aman Gupta 159c651f57 qwen4exp: fix tests (#29819) 2026-10-02 09:12:25 +08:00
Johannes Gäßler 78e2964c23 llama: refer to segment documentation [no ci] (#29074) 2026-10-01 21:52:40 +02:00
Georgi Gerganov 81e39ad343 llama : clamp kpool re-pool bound to existing pools (#29805)
* tests : simplify function signature

* llama : clamp kpool re-pool bound to existing pools

The n_tokens/kpool + n_seqs_unq bound on n_new_g overshoots when a batch
fills the whole cache: n_ctx tokens complete exactly n_ctx/kpool pools, so
the +1 pads new_pool_idxs/new_pool_rep one entry past n_pool_real. Graph
reserve only covers n_pool_real entries, so the first full-context decode
builds bigger tensors than reserved and ggml-alloc demands a graph
reallocation (abort under GGML_SCHED_DEBUG_REALLOC=1).

Clamp the bound to n_pool_real: a ubatch can never mark more pools than
the cache holds, and reserve's n_pool_max already covers that.

Assisted-by: pi:llama.cpp/MiMo-V2.6-Flash-MOPD

* cont : cap to n_pool_max
2026-10-01 19:55:34 +03:00
Georgi Gerganov 869034b4bb llama : fix invalid assert in recurrent memory (#29799) 2026-10-01 14:49:03 +03:00
ynankaniandJohannes Gäßler b56f34ab13 CUDA: Handle compute type for NVFP4 on cublass path (#29173)
* CUDA: Handle compute type for NVFP4 on cublass path

Signed-off-by: ynankani <ynankani@nvidia.com>

* Use BF16 compute type for quantized models if HW allows

Signed-off-by: ynankani <ynankani@nvidia.com>

* Set acc prec to bf16 for nvfp4 as it needs atleast bf16 range

Signed-off-by: ynankani <ynankani@nvidia.com>

* Update ggml/src/ggml-cuda/ggml-cuda.cu

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* preserve op_params for per-expert matmul

Signed-off-by: ynankani <ynankani@nvidia.com>

---------

Signed-off-by: ynankani <ynankani@nvidia.com>
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2026-10-01 16:53:52 +05:30
Aman GuptaandGeorgi Gerganov c061df1983 Qwen4Exp: add MTP (#29761)
* Qwen4Exp: add MTP

* remove has_state member, check via ctx_bufs being non-empty

* consistent naming + less verbose comments

* cont : clean-up recurrent memory

* cont : clean-up comments

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-10-01 14:13:27 +03:00
Aman Gupta 66e0c17ee1 llama: fix qwen4exp (#29751)
* llama: fix qwen4exp

* qwen4exp: keep kq_mask input the same shape
2026-10-01 14:13:27 +03:00
Pranesh GonegandlaandPranesh Gonegandla 32dd62ee6d llama-mmap : avoid a second full-size copy of each tensor with direct-io (#29749)
Assisted-by: Claude

Co-authored-by: Pranesh Gonegandla <pgonegandla@nvidia.com>
2026-10-01 09:44:31 +02:00
Toki NasinandSigbjørn Skjæret db33d3cb89 vocab : honor BOS/EOS settings for PLaMo-2 and PLaMo-3 (#29734)
* vocab : honor BOS/EOS settings for PLaMo-2 and PLaMo-3

The original tokenizer configs for PLaMo-2 and PLaMo-3 have
`add_bos_token: true` and `add_eos_token: false` , but
_set_vocab_plamo() did not write the BOS/EOS metadata. The
PLAMO2 tokenizer path also ignored add_bos/add_eos during
tokenization.

Write the settings from tokenizer_config.json and honor them in
the PLAMO2 tokenization path. GGUFs without these keys keep the
previous behavior.

* Update conversion/base.py

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
2026-10-01 08:31:42 +03:00
Kevin Hopper 10f340d1a2 model : re-enable -sm tensor for qwen4exp (#28569)
#27941 disabled -sm tensor for qwen4exp because test-llama-archs asserted on the
Meta device once the fixture carried a PLE layer:
GGML_ASSERT(ggml_backend_buffer_is_meta(tensor->buffer)) at ggml-backend-meta.cpp:476.

With host-resident embeddings the PLE gather is a CPU node and hc_init (the REPEAT
that fans the embedding out to the hc streams) was first reached through layer 0's
PLE path, after that gather. ggml_backend_sched_split_graph pass 2 expands a device
assignment upwards only until it meets a CPU node, so the REPEAT stayed on the CPU
and the later reshape of hc_init inside the meta split viewed a host-resident node.

Expanding hc_init right after it is built puts the REPEAT directly before the first
device node, where pass 2 assigns it; the embedding reshape stays in the CPU split
and is copied in as a split input, as in deepseek4.
2026-10-01 08:16:49 +03:00
Hrishith Thadicherla 4453b535fd llama : preserve original batch order for speculative decoding layer inputs (#29019)
* llama: preserve original batch order for layer inputs

Assisted-by: Codex

* tests: cover layer-input order across KV layouts

Assisted-by: Codex

* tests: exercise layer-input ordering on CUDA devices

Assisted-by: Codex

* llama: make layer input reordering compatible with tensor split

Copy each microbatch tensor from offset zero and restore original row order after synchronization. Extend the layer-input regression to cover tensor split and repeated reads and decodes.

Assisted-by: Codex

* llama: restore token order for unmasked NextN embeddings

Use the original-token mapping for unmasked NextN rows, including when
layer-input capture is disabled. Keep masked NextN rows on the logits
output mapping and preserve offset-zero tensor copies.

Extend the existing regression to cover NextN alone, combined layer
capture, and masked outputs with repeated decodes and getters.

Validation: all 256 CPU/CUDA/tensor configurations pass. Qwen3.8-27B
Q4_K_M MTP completes MT-Bench at concurrency 16 before and after.

Assisted-by: Codex

* ggml: fix WebGPU reservation and OpenVINO hidden-state capture

Reserve WebGPU vector attention scratch across batch sizes and refresh reservations when NextN capture settings change. Preserve requested OpenVINO outputs, dynamic shapes, sequence counts, and current graph bindings.

Extend existing WebGPU regression coverage and enable strict allocation checks.

Assisted-by: Codex

* llama: defer regression test and backend fixes to follow-ups

Keep this PR focused on restoring token order for layer inputs and unmasked NextN embeddings. Remove the added regression test, OpenVINO and WebGPU changes, and the separate NextN reservation change.

Assisted-by: Codex

* llama: keep n_embd declaration in its original position

Assisted-by: Codex

* llama : pass token count to layer input extraction

Assisted-by: Codex

* llama : name original batch indices batch_idxs

Assisted-by: Codex

* llama : name extracted embedding indices embd_batch_idxs

Assisted-by: Codex

* llama : tag target embedding reordering

Assisted-by: Codex

* llama : tag extraction and name the index capture flag

Assisted-by: Codex
2026-09-30 21:17:40 +03:00
Xuan-Son Nguyen 81ff93ea1d llama: properly handle KV on training (#28520)
* llama: properly handle KV on training

* improve
2026-09-30 18:09:08 +02:00
Pascal 05af0d2b13 glm5-next: give dead indexer slots unique scatter rows (#29745)
The sparse indexer mask is built with a set_rows scatter. Padded pools,
absent sequences and missing tail cells all pointed to the same n_kv
sentinel row, and invisible pools picked by top_k to fill the selection
overlap the tail cells of the token, so several CPU threads wrote the
same element (ThreadSanitizer data race in the sanitize CI).

Allocate the slot mask for both selection paths and route every dead
slot to its own dump row n_kv + slot. Live slots address disjoint cells,
so the scatter indices of a token are unique.
2026-09-30 17:10:48 +02:00
Georgi GerganovandSigbjørn Skjæret 22bdcc4cdd mimo : support dflash (convert + feature extraction) (#29650)
* convert : update to support dflash

* cont : fix

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
2026-09-30 17:04:35 +03:00
185103dcf5 llama: llama_prefetch_rows (#29599)
* llama: llama_prefetch_rows

* llama: support row prefetch on Windows

Apply the Windows port contributed by @praneshgo unchanged.

Source: https://github.com/ggml-org/llama.cpp/pull/29599#issuecomment-5887721014

* avoid exposing llama-mmap in model code, route via llama-impl

* add windows check, only prefetch in lazy mode

* cont : clean-up

* cont : fix build

* cont : clarify padding token for gemma4

---------

Co-authored-by: Pranesh Gonegandla <pranesh.iitp@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-09-30 20:27:18 +08:00