Commit graph

9013 commits

Author SHA1 Message Date
Aldehir Rojas
e48034dfc9
common : determine generation prompt using longest common prefix (#22657) 2026-05-04 00:18:23 +02:00
Julien Denize
048a490f76
convert : Mistral format yarn apply_scale support (#22612)
* [BUGFIX] Mistral format apply_scale support.

* Update convert_hf_to_gguf.py

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

* fix misunderstood boolean parameters

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-05-03 21:51:21 +02:00
JM Robles
db44417b02
convert : apply Q/K RoPE permutation in NVFP4 repack path (#22611)
Llama-architecture q_proj/k_proj weights need an axis-0 row permutation
to match GGML's RoPE convention. The BF16 path applies this in
LlamaModel.modify_tensors via LlamaModel.permute, but the NVFP4 path
bypasses modify_tensors and writes weights directly through
ModelBase._repack_nvfp4. Without the permutation, attention heads end
up scrambled at inference and the model produces gibberish.

This change overrides _repack_nvfp4 on LlamaModel and applies the same
permutation to both the nibble-packed weight and the per-block scale
before delegating to ModelBase._repack_nvfp4 via super(). Reuses the
existing LlamaModel.permute static helper and respects the existing
undo_permute flag, so subclasses (Mistral, Granite, Llama4, etc.)
inherit the fix automatically.

Verified on TinyLlama-1.1B reproducer: perplexity drops from 4419
(gibberish) to 43.9, matching the BF16-dequantized baseline (44.0).
Also verified end-to-end on ALIA-40b-instruct-2601 (BSC, Llama
architecture) with multilingual generation in Spanish/Catalan/Basque/
Galician all coherent with the fix applied.

Co-authored-by: Chema <chema@montevive.ai>
2026-05-03 18:22:00 +03:00
lucy
d05fe1d7da
fix: CUDA device PCI bus ID de-dupe OOMing (ignoring other 3 gpus entirely) (#22533)
* fix: CUDA device PCI bus ID detection for multi-GPU de-dupe

* HIP, MUSA macros

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2026-05-02 22:19:25 +02:00
Georgi Gerganov
0754b7b6fe
server : avoid checkpoint data host copies (#22558)
* server : avoid checkpoint data host copies

* llama : refactor llama_io_read_i
2026-05-02 18:03:25 +03:00
JusteLeo
09294365a9
ggml-virtgpu: fix circular dependency in headers (#22557) 2026-05-02 21:28:50 +08:00
Csaba Kecskemeti
63d93d1733
convert : disable uint types (#18908)
Some checks failed
Check Pre-Tokenizer Hashes / pre-tokenizer-hashes (push) Has been cancelled
Python check requirements.txt / check-requirements (push) Has been cancelled
Python Type-Check / python type-check (push) Has been cancelled
2026-05-02 09:05:59 +03:00
Shawn Gu
c5a3bc39b1
opencl: Adreno optimization for MoE - MxFP4 (#22301)
* MoE Mxfp4 CLC kernel added, router reorder on GPU

* Pass test-backend-ops for MoE mxfp4 Adreno CLC

* remove putenv in llama-model.cpp

* fix indent style and whitespace

* opencl: remove unnecessary headers

* opencl: do not save cl_program objects

* opencl: remove unnecessary assert

* fix precision issue

---------

Co-authored-by: Li He <lih@qti.qualcomm.com>
2026-05-01 23:02:24 -07:00
Johannes Gäßler
9dbb372610
Github: update issue templates (#22594) 2026-05-02 07:56:13 +02:00
Georgi Gerganov
228e836344 sync : ggml 2026-05-02 08:55:29 +03:00
Georgi Gerganov
ed23489f42 ggml : bump version to 0.10.2 (ggml/1474) 2026-05-02 08:55:29 +03:00
Georgi Gerganov
457e2288c9 sync : ggml 2026-05-02 07:22:35 +03:00
Georgi Gerganov
e8ec7ab058 ggml : try fix win32 build (whisper/0) 2026-05-02 07:22:35 +03:00
Yiwei Shao
1a03cf47f6
hexagon: hmx flash attention (#22347)
* hmx: extract shared interleave headers and unify matmul batched

* hmx: add HMX-accelerated flash attention for prefill

* hmx: replace asm wrappers with Q6_ intrinsics in hmx-utils.h

Switches three single-instruction helpers from inline asm to the matching
Q6_ intrinsics, matching the style established by aizip f8737609a and used
by the upstream PR #21554 hmx-matmul-ops.c rewrite:

  hmx_set_output_scales       asm "bias=mxmem2"  -> Q6_bias_mxmem2_A
  hmx_load_tile_pair_fp16     asm packet         -> Q6_activation_hf_mxmem_RR
                                                    + Q6_weight_hf_mxmem_RR
  hmx_consume_accumulator_fp16 asm "mxmem=acc"   -> Q6_mxmem_AR_after_hf

hmx_load_tiles_fp16 stays on inline asm: it uses ":deep" activation
streaming, and the mixed Q6_activation_hf_mxmem_RR_deep + non-deep
Q6_weight_hf_mxmem_RR pair fails the HMX backend constraint check
("activate weight pair (1) exceeds limit (1)"). The asm bundle keeps
both halves in one VLIW packet and avoids the diagnostic.

Functionally equivalent — same instructions emitted; the Q6_ intrinsics
just give the compiler more visibility for scheduling.

* hmx: drop the duplicate interleave_fp16_weight_chunk_to_tiles

* hmx:  apply upstream optimization to hmx-flash-attn-ops.c
apply restrict, __builtin_assume, and pointer accumulation to the three HMX workers (qk_dot, o_update, o_norm) and the matching inline HMX loops in op_hmx_flash_attn_ext.

* hmx: unify interleave helper

* hmx: multi-thread Q load / O store and enable prefill FA dispatch

Extract inline Q-load and O-store loops into worker_pool-parallel helpers
(fa_phase_q_load, fa_phase_o_store) so HVX threads split the F32↔F16
conversion work across row ranges.  Also relax the softmax threading
gate from n_row_vec_cnt >= n_threads to >= 2, which was unnecessarily
forcing single-thread fallback when n_rows_g < 512.

On the dispatch side, remove the ne[2] != 1 guard that blocked multi-head
(prefill) FA from reaching the HTP backend — GQA is already handled
internally by both the HMX and HVX flash-attention paths.

* hmx: relax matmul pipeline gate to cover k > n shapes (e.g. FFN_down)

* hmx: optimize FA softmax mask phase (no-ALiBi fast path + GQA dedup)

* hmx: Add an asm memory clobber at the phase boundary to prevent reorder bug

* [experimental]: fp16 softmax (EXP2_HF) to accelerate fa

Bake log2(e) into qk_scale and use hvx_exp2_hf directly for P and m_diff
(base-2 consistent, matches htp-ops-lib). ~22 ALU ops for 64 lanes vs
~44 for the F32 round-trip path.

* hmx flash-attn: refine cost model coefficients based on profiling data

* hmx flash-attn: replace asm clobber with targeted volatile reads on vtcm_d_tiles

* hmx flash-attn: fix prefill correctness (dst indexing, softmax reduce, V stride)

* hmx flash-attn: fix p_tiles dual-tile OOB race; enable MT + pipeline

* hmx flash-attn: preserve additive mask bias in no-ALiBi fast path

The no-ALiBi fast path (max_bias==0) was skipping mask add entirely on
the assumption that mask values are only {0, -inf}.  This is wrong when
the mask carries additive positional bias — those terms were silently
dropped.  Keep the slope-mul skip (slope≡1.0) but add mask back so the
bias survives; vmux still clamps below -16 to -inf.

Also add HMX FA coverage to test-backend-ops: prefill shapes (nb=64,
nb=32) × {mask on/off} × {ALiBi on/off} × {softcap on/off}, F16 KV,
hs ∈ {64, 128}.

* hmx: fix softcap+EXP2_HF interaction, tighten matmul pipeline gate, add FA tests

- flash-attn: when EXP2_HF is on AND logit_softcap is active, fold
  log2(e) into the post-tanh multiplier (v_cap) instead of pre-baking
  it into qk_scale.  Pre-baking shifted the tanh knee from x≈c to
  x≈c/log2(e) and produced numerically wrong softcapped outputs
  whenever both knobs were enabled.
- flash-attn softmax (fa_softmax_thread): replace the union+memcpy
  scalar extract pattern with HVX vmux-based per-row accumulators on
  rowmax/rowsum.  Add hvx_vec_get_f16 helper in hvx-base.h.  Functional
  parity, less scalar code, clearer hf/qf16 lane-format contract.
- matmul (hmx_mat_mul_permuted_qk_0_d16a32): pick pipeline vs sequential
  layout based on whether the chunker actually yields >=2 n-chunks,
  instead of the static (m>=128 && n>=256) gate.  Avoids paying for
  output double-buffer + worker dispatch when there is no HMX/HVX
  overlap to gain (e.g. shapes that collapse to one n-chunk).
- tests: add HMX flash-attention coverage over the
  {mask, ALiBi (max_bias), logit_softcap} cross-product for the prefill
  path — head_dim 64/128, GQA 4×4, kv=512/nb=64 plus a kv=113/nb=32
  non-aligned case.

* [Help Wanted]: refactor D matrix computation into separate function for clarity and maintainability

* format code

* hexagon: looks like -O3 is causing issues with the large code base, switch to -O2 and -flto instead

* hexagon: use hex_ prefix for swap_ptr

* hexagon: move vtcm_seq_alloc into vtcm-utils.h

More vtcm allocator updates are coming so it makes sense to start the separate hdr for it.

* hmx-utils: add hmx_prefix for layout converters

* hmx-mm: move main hmx_mm functions to the end, remove unused fwd decls, etc

* hmx-mm: remove unused qweight_fetch_task_state_t and minor alignment fixes

* hmx-fa: minor alignment fixes

* hmx-fa: move hmx_flash_atten into hmx-ops.h

* hmx-fa: remove redundant workpool pointer in the hmx_fa_ctx, plus minor alignment updates

* hmx-fa: minor alignment and simplifications

* hexagon: move FA_EXP_F16 option to hostside CMake file

* hmx-fa: use hvx_vec_splat_f16 instead of fp16_to_bits

* hmx-fa: add hvx_splat_u16/u8 and use that in the fa instead custom hvx_fill

* hmx-fa: some more alignment updates in the core fa function

* hmx-fa: keep slopes in vtcm in fp16

Saves malloc/free and removes the need for float -> fp16 downcast on every use.

* hexagon: consistent noinline usage (after static)

* hex-hmx: consistent use FARF_HIGH to enable debug output

* hmx-utils: no need for always_inline attr

* hex-hmx: consistent noinline usage (static noinline ...)

* hex-hmx: simplify init_col_scales

* hexagon: fix editorconfig errors

* hmx-mm: minor alignment fixes

---------

Co-authored-by: Max Krasnyansky <maxk@qti.qualcomm.com>
2026-05-01 20:29:13 -07:00
ddh0
b97ebdc98f
llama-quant : fix --tensor-type when default qtype is overriden (#22572)
fix #22544 (my fault!)

Credit to @Anai-Guo, ref #22559 - since that one was closed due to the
new contributor policy I am taking the liberty of re-submitting that PR
here.
2026-05-01 19:55:55 +02:00
Aparna M P
2098fd6169
hexagon: enable non-contiguous row tensor support for unary ops (#22574) 2026-05-01 10:09:23 -07:00
Aleksander Grygier
ab6120cde5
webui: Spring Cleaning Refactor v1 (#22505)
* wip: server_tools

* feat: Integrate with `/tools` endpoint

* feat: Builtin + MCP + JSON Schema Tools WIP

* refactor

* displayName -> display_name

* snake_case everywhere

* rm redundant field

* feat: Improvements

* chore: update webui build output

* refactor: Updates after server updates

* chore: update webui build output

* change arg to --tools all

* feat: UI improvements

* chore: update webui build output

* add readme mention

* llama-gen-docs

* chore: update webui build output

* chore: update webui build output

* chore: update webui build output

* feat: Reorganize settings sections

* feat: Separate dialogs for MCP Servers Settings and Import/Export

* feat: WIP

* feat: WIP

* feat: WIP

* feat: WIP

* feat: WIP

* feat: WIP

* WIP on allozaur/20677-webui-server-tools

* feat: UI improvements

* chore: Update package lock

* chore: Run `npm audit fix`

* feat: UI WIP

* feat: UI

* refactor: Desktop Icon Strip DRY

* feat: Cleaner rendering and transition for ChatScreen

* feat: UI improvements

* feat: UI improvement

* feat: Remove MCP Server "enable" switch from Tools submenu

* chore: Run `npm audit fix`

* feat: WIP

* feat: Logic improvements

* refactor: Cleanup

* refactor: DRY

* test: Fix Chat Sidebar UI Tests

* chore: Update package lock

* refactor: Cleanup

* feat: Chat Message Action Card with Continue and Permission flow implementations

* feat: Add agentic steering messages, draft messages and improve chat UX

* fix: Search results UI

* test: Fix unit test

* feat: UI/UX improvements

* refactor: Simplify `useToolsPanel` access in components

* feat: Implement Processing Info Context API

* feat: Implement 'Go back to chat' functionality for settings

* feat: Enhance MCP Server management in Chat Form Attachments

* style: Minor UI and branding adjustments

* chore: Update webui static build output

* chore: Formatting, linting & type checks

* feat: Draft messages logic

* feat: UI improvements

* feat: Steering Messages improvements

* refactor: Cleanup

* refactor: Cleanup

* feat: Improve UI

* refactor: Settings navigation hook

* refactor: DRY code

* refactor: DRY ChatMessageUser UI components

* refactor: Desktop Icon Strip DRY

* refactor: Tools & permissions

* fix: Navigation condition

* refactor: Cleanup

* refactor: Cleanup

* refactor: Cleanup

* fix: preserve reasoning_content in agentic flow

* refactor: Storybook cleanup

* refactor: isInViewport util function

* refactor: Rename globally `onClick` to `onclick`

* chore: `npm audit fix`

* refactor: Action Icon usage

* refactor: Naming

* refactor: JS in `class` directive

* refactor: Chat components cleanup WIP

* refactor: Components structure

* refactor: Cleanup WIP

* feat: New ChatAttachmentsPreview component

* feat: UI improvements

* feat: UI improvements

* refactor: Cleanup

* refactor: ChatAttachmentsPreview UI/UX

* refactor: Remove dead code

* refactor: Cleanup

* fix: Model Name aliases displaying

* feat: Shortcut improvements

* refactor: Chat Message

* feat: Move Import/Export to settings

* refactor: Cleanup

* refactor: Cleanup

* refactor: Cleanup

* refactor: Cleanup

---------

Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
2026-05-01 18:36:29 +02:00
Masashi Yoshimura
c3c1505392
ggml-webgpu: Fix vectorized handling in mul-mat and mul-mat-id (#22578)
* Fix vectorized condition of mul-mat-fast pipeline and add vectorized variant to mul-mat-id

* Apply suggestion from @CISC

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

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-05-01 07:55:01 -07:00
Jeff Bolz
05e141a6b3
vulkan: Support asymmetric FA in coopmat2 path (#21753)
* vulkan: Support asymmetric FA in coopmat2 path

There has been some recent interest/experimentation with mixed quantization
types for FA. I had originally designed the cm2 FA shader with this in mind
(because I didn't realize it wasn't supported at the time!), this change
adds the missing pieces and enables it.

Also support Q1_0 since people have been trying that out (seems crazy, but
who knows).

We should be able to do similar things in the coopmat1/scalar path, but
there's another change open against the scalar path and I don't want to
conflict.

* reorder cases
2026-05-01 15:28:32 +02:00
Chen Yuan
aab68217b7
ggml-webgpu: add the upscale shader (#22419)
* shader(upscale): add the upscale shader with nearest, bilinear and bicubic implementations

* shader(upscale): use macro
2026-04-30 22:22:18 -07:00
Masashi Yoshimura
a95a11e5b8
ggml-webgpu: Improve performance of mat-vec and mat-mat for MUL_MAT_ID (#22464)
* Add mat-vec fast path of MUL_MAT_ID.

* Add shared accumulation vec logic and the other types supports.

* Add i-quant mat-mat for MUL_MAT_ID and fix some parts

* Remove n_experts from shader_lib_context.
2026-04-30 14:19:10 -07:00
Reese Levine
5cbfb18075
Update llama-mmap to use ftello/fseeko (#22497)
* Update llama-mmap to work with 32-bit wasm and >2GB models

* Update to gguf.cpp style
2026-04-30 14:17:52 -07:00
Adrien Gallouët
beb42fffa4
common : check for null getpwuid in hf-cache (#22550)
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-04-30 21:32:41 +02:00
Ruben Ortlam
660b1b4bdc
vulkan: add get/set tensor 2d functions (#22514)
Some checks failed
Check Pre-Tokenizer Hashes / pre-tokenizer-hashes (push) Has been cancelled
Python check requirements.txt / check-requirements (push) Has been cancelled
Python Type-Check / python type-check (push) Has been cancelled
* vulkan: add get/set_tensor_2d functions

* fix backend interface comments

* Update ggml/src/ggml-metal/ggml-metal.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-04-30 17:37:13 +02:00
Ben Guidarelli
c20c44514a
spec: fix argument typo (#22552) 2026-04-30 17:32:32 +03:00
Sigbjørn Skjæret
6118c043b1
ci : bump ty to 0.0.33 (#22535)
* bump ty to 0.0.33

* update typings
2026-04-30 16:15:54 +03:00
Adrien Gallouët
5f0ab726f7
vendor : update cpp-httplib to 0.43.2 (#22548)
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-04-30 15:04:39 +02:00
Johannes Gäßler
e82aaf2587
CUDA: fix tile FA kernel on Pascal (#22541) 2026-04-30 13:04:50 +02:00
Georgi Gerganov
27aef3dd91
scripts : add wc2wt.sh - create worktree from current HEAD (#22513)
* scripts : add wc2wt.sh - create worktree from current HEAD

Add a script to create a git worktree on a new branch from the current
HEAD. Similar to pr2wt.sh but for local development branches instead of
PRs.

Usage:
  ./scripts/wc2wt.sh gg/new-feature
  ./scripts/wc2wt.sh gg/new-feature "bash -l"

Assisted-by: llama.cpp:local pi

* cont : no need to try to delete the branch
2026-04-30 09:20:26 +03:00
Rithik Sharma
45155597aa
add fast matmul iquants (#22504) 2026-04-29 22:58:32 -07:00
Georgi Gerganov
80afa33aad
spec : fix draft model checkpoints (#22521)
* spec : fix draft model checkpoints

* cont : clean-up

* cont : gate the ngram-mod reset warning behind verbose flag
2026-04-30 08:32:18 +03:00
Peter Sideris
b42c7fa5b8
spec : fix vocab compat checks in spec example (#22426)
* port #22358 PR to examples/speculative/speculative.cpp
* use vocab_[tgt,dft] instead of ctx_[tgt,dft] when logging on draft
  model / target model vocabulary mismatch

Co-authored-by: Petros Sideris <petros.sideris@nokia.com>
2026-04-30 08:18:25 +03:00
Aldehir Rojas
d77599234e
common : do not pass prompt tokens to reasoning budget sampler (#22488) 2026-04-29 14:10:58 -05:00
Max Krasnyansky
41a63be28e
hexagon: make vmem and buffer-size configurable (#22487)
* hexagon: allow host to set max vmem size

We use a sane default but it's helpful to allow for an override if needed.

* hexagon: add support for measuring vmem space and move pinned mmaping management to host

* hexagon: update vmem checks to use uint64

* hexagon: bump op buffers to 16 (matches max mmaps)

* hexagon: bump default vmem to 3.2GB

* hexagon: add support for autodetecting vmem space and some logging cleanup in that area

* hexagon: fix whitespace warnings

* Update scripts/snapdragon/adb/run-cli.sh

Co-authored-by: Pascal <admin@serveurperso.com>

* hex-adb: fix run-completion script

---------

Co-authored-by: Pascal <admin@serveurperso.com>
2026-04-29 11:51:21 -07:00
Anav Prasad
098705a29e
CUDA: fuse SSM_CONV + ADD(bias) + SILU (#22478) 2026-04-30 02:39:56 +08:00
Georgi Gerganov
683c5acb90
spec : disacard last drafted token with low prob (#22506)
Some checks are pending
Check Pre-Tokenizer Hashes / pre-tokenizer-hashes (push) Waiting to run
Python check requirements.txt / check-requirements (push) Waiting to run
Python Type-Check / python type-check (push) Waiting to run
2026-04-29 17:00:00 +03:00
Georgi Gerganov
b1d5f5b449 sync : ggml 2026-04-29 16:43:47 +03:00
Georgi Gerganov
4b221b7f1e ggml : bump version to 0.10.1 (ggml/1469) 2026-04-29 16:43:47 +03:00
Pascal
59237bfbbc
webui: fix slow mic stop and WAV encode (#22480)
* webui: instant mic stop, race-free recorder restart

* webui: faster WAV PCM encode via hoisted channels and Int16Array

* chore: update webui build output

* webui: drop setTimeout(0) hack and harden cancelRecording

* chore: update webui build output
2026-04-29 12:58:35 +02:00
shalinib-ibm
1cbc846eba
ggml-cpu : disable tiled matmul on AIX to fix page boundary segfault (#22293)
* ggml-cpu : disable tiled matmul on AIX to fix page boundary segfault

vec_xst operations in the tiled path crash on AIX when writing
near 4KB page boundaries due to strict memory protection. Fall
back to mnpack implementation on AIX for stable execution.

Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>

* Update ggml/src/ggml-cpu/llamafile/sgemm.cpp

Co-authored-by: Aaron Teo <taronaeo@gmail.com>

* Update sgemm.cpp

* Update sgemm.cpp

---------

Signed-off-by: Shalini Salomi Bodapati <Shalini.Salomi.Bodapati@ibm.com>
Co-authored-by: Aaron Teo <taronaeo@gmail.com>
2026-04-29 13:32:40 +03:00
Aman Gupta
3142f1dbb9
ggml-cuda: refactor fusion code (#22468)
* ggml-cuda: refactor fusion code

* apply formatting + make env variable truthy
2026-04-29 16:19:33 +08:00
qiurui144
b5c4227dc6
ggml-cpu: cmake: append xsmtvdotii march for SpacemiT IME (#22317)
* ggml-cpu: cmake: append xsmtvdotii march for SpacemiT IME

When GGML_CPU_RISCV64_SPACEMIT=ON is set, ime1_kernels.cpp contains
inline asm for the vmadot family which requires the xsmtvdotii custom
extension.(problem can see in some blogs and make sure in K3 platform)
The current CMakeLists does not include xsmtvdotii, so any toolchain
that honours the explicit -march (tested with SpacemiT GCC 15.2) fails
at the assembler stage:

  Error: unrecognized opcode `vmadot v16,v14,v0',
         extension `xsmtvdotii' required

Append _xsmtvdotii to MARCH_STR when GGML_CPU_RISCV64_SPACEMIT is
enabled so the IME path can actually build with a capable toolchain.
No effect on builds that leave GGML_CPU_RISCV64_SPACEMIT off.

toolchain from https://www.spacemit.com/community/resources-download/Tools

* Update ggml/src/ggml-cpu/CMakeLists.txt

Co-authored-by: alex-spacemit <jinghui.huang@spacemit.com>

---------

Co-authored-by: alex-spacemit <jinghui.huang@spacemit.com>
2026-04-29 10:59:21 +03:00
Reese Levine
d6a5094004
ggml-webgpu: Fix bug in FlashAttention support check (#22492)
* Fix flashattention support check for devices that don't support subgroups

* set path to none if kv_tile doesn't fit
2026-04-29 10:59:00 +03:00
Masato Nakasaka
7b95ea5d11
common: Intentionally leak logger instance to fix hanging on Windows (#22273)
* Changed to leak logger singleton to prevent hanging on Windows

* Fix comment

* Stopped using static vector

Using std::vector will cause g_col to be released before the logger thread exits, causing the logger thread to touch freed memory causing a crash

* Change so all logs are output before exit

* Added debug logging

* added more logging

* Added logging

* Explicitly free logger to avoid hanging on Win

* Reverted to leak logger instance again

* Removed debug log and fixed comment

* Fixed comment

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-04-29 10:58:43 +03:00
hrushitfujitsu
bdc9c743a5
ggml : add sve tuned code for gemm_q8_0_4x8_q8_0() kernel (#21916)
* Added sve tuned code for gemm_q8_0_4x8_q8_0() kernel

* Change arrays to static const in repack.cpp

---------

Co-authored-by: Vithulep <prashant.vithule@fujitsu.com>
2026-04-29 10:57:37 +03:00
Johannes Gäßler
739393beeb
TP: fix delayed AllReduce + zero-sized slices (#22489) 2026-04-29 08:55:07 +02:00
Michael Wand
fc2b0053ff
ggml-cuda: Repost of 21896: Blackwell native NVFP4 support (#22196) 2026-04-29 06:47:42 +08:00
lnigam
7b8443ac78
ggml-cuda: add flash-attn support for DKQ=320/DV=256 with ncols2=32 (… (#22286)
* ggml-cuda: add flash-attn support for DKQ=320/DV=256 with ncols2=32 (GQA=32)

Adds MMA-f16 and tile kernel configs, dispatch logic, template instances,
and tile .cu file for Mistral Small 4 (head sizes 320/256), restricting to
ncols2=32 to support GQA ratio 32 only.

* Adding check to return BEST_FATTN_KERNEL_NONE in case GQA!=32

* Apply suggestions from code review

Address review comments

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

* Address review comments and making kernel config default to DQK=512, DV=512 instead of DQK=256,DV=256

* Fixed bug with sinks=1, with ncols=32, there are two warp-groups created but sinks index is same(0,...,15) for both the groups hence with sinks=1, output is not matching with CPU output. Added sink_base which will be base index for each warp_group (threadIdx.y / np)

* Apply suggestions from code review

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

* Update ggml/src/ggml-cuda/template-instances/generate_cu_files.py

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

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2026-04-28 21:37:35 +02:00
Daniel Bevenius
5d56effdee
convert : add support for Nemotron Nano 3 Omni (#22481)
This commit adds support for NVIDIA Nemotron Nano 3 Omni model enabling
this model to be converted to GGUF.
2026-04-28 19:17:57 +02:00
Jillis ter Hove
52e5f0a5c1
common : re-arm reasoning budget after DONE on new <think> (#22323)
DONE state absorbs all tokens including a new start tag, causing any think blocks after the first to run unbudgeted. Observed on unsloth/Qwen3.6-27B-GGUF which interleaves multiple <think> blocks per response.

Fixed by advancing start_matcher in DONE branch and re-arming to COUNTING with a fresh budget on match. Adds regression test (test-reasoning-budget: test 6).
2026-04-28 19:15:36 +02:00