]>
git.djapps.eu Git - pkg/ggml/sources/llama.cpp/log
TianHao324 [Tue, 2 Dec 2025 12:35:23 +0000 (20:35 +0800)]
CANN: Disable Ger operator of OUT_PROD on 310p device (#17563)
Daniel Bevenius [Tue, 2 Dec 2025 11:52:45 +0000 (12:52 +0100)]
ggml : remove redundant n_copies check when setting input/output (#17612)
This commit removes a redundant check for sched->n_copies > 1 when
setting input and output flags on tensor copies in
ggml_backend_sched_split_graph.
The motivation for this change is to clarify the code as the outer if
statement already performs this check.
Eric Curtin [Tue, 2 Dec 2025 11:18:15 +0000 (11:18 +0000)]
codeowners : remove ericcurtin (#17658)
Taking a break from llama.cpp . I wasn't around at the start of llama.cpp
but I want to thank @ggerganov and @slaren for creating a neat community
here.
Signed-off-by: Eric Curtin <redacted>
Adrien Gallouët [Tue, 2 Dec 2025 11:05:38 +0000 (12:05 +0100)]
llama : fix signed comparison warning on FreeBSD (#17497)
This ensures correct RLIM_INFINITY handling and compatibility on all platforms (32/64-bit).
warning: comparison of integers of different signs: 'rlim_t' (aka 'long') and 'size_t' (aka 'unsigned long') [-Wsign-compare]
488 | if (suggest && (lock_limit.rlim_max > lock_limit.rlim_cur + size)) {
| ~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Adrien Gallouët <redacted>
Xuan-Son Nguyen [Tue, 2 Dec 2025 10:48:31 +0000 (11:48 +0100)]
convert: add error message for mistral3 quantized weight (#17686)
Xuan-Son Nguyen [Tue, 2 Dec 2025 10:38:57 +0000 (11:38 +0100)]
server: remove default "gpt-3.5-turbo" model name (#17668)
* server: remove default "gpt-3.5-turbo" model name
* do not reflect back model name from request
* fix test
senhtry [Tue, 2 Dec 2025 10:18:39 +0000 (18:18 +0800)]
server: fixing naming conflict res_error in server-models.cpp (#17679)
Xuan-Son Nguyen [Tue, 2 Dec 2025 09:25:11 +0000 (10:25 +0100)]
server: explicitly set exec path when create new instance (#17669)
* Revert "rm unused fn"
This reverts commit
f2dbe9c087ac8776cd41f747a8b0ddf448c8399a .
* server: explicitly set exec path when create new instance
* put back TODO
* only call get_server_exec_path() once
* add fallback logic
Adrien Gallouët [Tue, 2 Dec 2025 09:15:01 +0000 (10:15 +0100)]
ci : skip winget update when not in ggml-org (#17465)
Prevent forks from generating daily failure notifications.
Signed-off-by: Adrien Gallouët <redacted>
Adrien Gallouët [Tue, 2 Dec 2025 08:41:26 +0000 (09:41 +0100)]
ggml : add fallback definition for HWCAP2_SVE2 (#17683)
This align with other HWCAP2 feature flags
See #17528
Signed-off-by: Adrien Gallouët <redacted>
Aleksander Grygier [Tue, 2 Dec 2025 08:20:57 +0000 (09:20 +0100)]
Add context info to server error (#17663)
* fix: Add context info to server error
* chore: update webui build output
Aman Gupta [Tue, 2 Dec 2025 04:36:31 +0000 (12:36 +0800)]
ggml-cuda: reorder only relevant nodes (#17639)
Aaron Teo [Tue, 2 Dec 2025 03:52:05 +0000 (11:52 +0800)]
release: fix duplicate libs, store symbolic links (#17299)
Neo Zhang Jianyu [Tue, 2 Dec 2025 00:56:46 +0000 (08:56 +0800)]
enhance argsort for UT (#17573)
Co-authored-by: Neo Zhang <redacted>
Piotr Wilkin (ilintar) [Mon, 1 Dec 2025 23:43:13 +0000 (00:43 +0100)]
Override SSM_A op for Qwen3 Next to reduce splits (#17587)
* Override SSM_A op for Qwen3 Next to reduce splits
* New tensor mapping SSM_A_NOSCAN for SSM_A used outside of OP_SSM_SCAN context.
* Update src/llama-model.cpp
Co-authored-by: Sigbjørn Skjæret <redacted>
* Update src/llama-model.cpp
Co-authored-by: Sigbjørn Skjæret <redacted>
---------
Co-authored-by: Sigbjørn Skjæret <redacted>
Jeff Bolz [Mon, 1 Dec 2025 21:26:21 +0000 (15:26 -0600)]
ops.md: update vulkan support (#17661)
Xuan-Son Nguyen [Mon, 1 Dec 2025 20:32:25 +0000 (21:32 +0100)]
mtmd: add mtmd_context_params::warmup option (#17652)
* mtmd: add mtmd_context_params::warmup option
* reuse the common_params::warmup
Gilad S. [Mon, 1 Dec 2025 20:21:13 +0000 (22:21 +0200)]
fix: llama arch implementation (#17665)
Xuan-Son Nguyen [Mon, 1 Dec 2025 18:41:04 +0000 (19:41 +0100)]
server: introduce API for serving / loading / unloading multiple models (#17470)
* server: add model management and proxy
* fix compile error
* does this fix windows?
* fix windows build
* use subprocess.h, better logging
* add test
* fix windows
* feat: Model/Router server architecture WIP
* more stable
* fix unsafe pointer
* also allow terminate loading model
* add is_active()
* refactor: Architecture improvements
* tmp apply upstream fix
* address most problems
* address thread safety issue
* address review comment
* add docs (first version)
* address review comment
* feat: Improved UX for model information, modality interactions etc
* chore: update webui build output
* refactor: Use only the message data `model` property for displaying model used info
* chore: update webui build output
* add --models-dir param
* feat: New Model Selection UX WIP
* chore: update webui build output
* feat: Add auto-mic setting
* feat: Attachments UX improvements
* implement LRU
* remove default model path
* better --models-dir
* add env for args
* address review comments
* fix compile
* refactor: Chat Form Submit component
* ad endpoint docs
* Merge remote-tracking branch 'webui/allozaur/server_model_management_v1_2' into xsn/server_model_maagement_v1_2
Co-authored-by: Aleksander <redacted>
* feat: Add copy to clipboard to model name in model info dialog
* feat: Model unavailable UI state for model selector
* feat: Chat Form Actions UI logic improvements
* feat: Auto-select model from last assistant response
* chore: update webui build output
* expose args and exit_code in API
* add note
* support extra_args on loading model
* allow reusing args if auto_load
* typo docs
* oai-compat /models endpoint
* cleaner
* address review comments
* feat: Use `model` property for displaying the `repo/model-name` naming format
* refactor: Attachments data
* chore: update webui build output
* refactor: Enum imports
* feat: Improve Model Selector responsiveness
* chore: update webui build output
* refactor: Cleanup
* refactor: Cleanup
* refactor: Formatters
* chore: update webui build output
* refactor: Copy To Clipboard Icon component
* chore: update webui build output
* refactor: Cleanup
* chore: update webui build output
* refactor: UI badges
* chore: update webui build output
* refactor: Cleanup
* refactor: Cleanup
* chore: update webui build output
* add --models-allow-extra-args for security
* nits
* add stdin_file
* fix merge
* fix: Retrieve lost setting after resolving merge conflict
* refactor: DatabaseStore -> DatabaseService
* refactor: Database, Conversations & Chat services + stores architecture improvements (WIP)
* refactor: Remove redundant settings
* refactor: Multi-model business logic WIP
* chore: update webui build output
* feat: Switching models logic for ChatForm or when regenerating messges + modality detection logic
* chore: update webui build output
* fix: Add `untrack` inside chat processing info data logic to prevent infinite effect
* fix: Regenerate
* feat: Remove redundant settigns + rearrange
* fix: Audio attachments
* refactor: Icons
* chore: update webui build output
* feat: Model management and selection features WIP
* chore: update webui build output
* refactor: Improve server properties management
* refactor: Icons
* chore: update webui build output
* feat: Improve model loading/unloading status updates
* chore: update webui build output
* refactor: Improve API header management via utility functions
* remove support for extra args
* set hf_repo/docker_repo as model alias when posible
* refactor: Remove ConversationsService
* refactor: Chat requests abort handling
* refactor: Server store
* tmp webui build
* refactor: Model modality handling
* chore: update webui build output
* refactor: Processing state reactivity
* fix: UI
* refactor: Services/Stores syntax + logic improvements
Refactors components to access stores directly instead of using exported getter functions.
This change centralizes store access and logic, simplifying component code and improving maintainability by reducing the number of exported functions and promoting direct store interaction.
Removes exported getter functions from `chat.svelte.ts`, `conversations.svelte.ts`, `models.svelte.ts` and `settings.svelte.ts`.
* refactor: Architecture cleanup
* feat: Improve statistic badges
* feat: Condition available models based on modality + better model loading strategy & UX
* docs: Architecture documentation
* feat: Update logic for PDF as Image
* add TODO for http client
* refactor: Enhance model info and attachment handling
* chore: update webui build output
* refactor: Components naming
* chore: update webui build output
* refactor: Cleanup
* refactor: DRY `getAttachmentDisplayItems` function + fix UI
* chore: update webui build output
* fix: Modality detection improvement for text-based PDF attachments
* refactor: Cleanup
* docs: Add info comment
* refactor: Cleanup
* re
* refactor: Cleanup
* refactor: Cleanup
* feat: Attachment logic & UI improvements
* refactor: Constants
* feat: Improve UI sidebar background color
* chore: update webui build output
* refactor: Utils imports + move types to `app.d.ts`
* test: Fix Storybook mocks
* chore: update webui build output
* test: Update Chat Form UI tests
* refactor: Tooltip Provider from core layout
* refactor: Tests to separate location
* decouple server_models from server_routes
* test: Move demo test to tests/server
* refactor: Remove redundant method
* chore: update webui build output
* also route anthropic endpoints
* fix duplicated arg
* fix invalid ptr to shutdown_handler
* server : minor
* rm unused fn
* add ?autoload=true|false query param
* refactor: Remove redundant code
* docs: Update README documentations + architecture & data flow diagrams
* fix: Disable autoload on calling server props for the model
* chore: update webui build output
* fix ubuntu build
* fix: Model status reactivity
* fix: Modality detection for MODEL mode
* chore: update webui build output
---------
Co-authored-by: Aleksander Grygier <redacted>
Co-authored-by: Georgi Gerganov <redacted>
Xuan-Son Nguyen [Mon, 1 Dec 2025 13:38:13 +0000 (14:38 +0100)]
common: improve verbosity level definitions (#17630)
* common: improve verbosity level definitions
* string_format
* update autogen docs
Xuan-Son Nguyen [Mon, 1 Dec 2025 11:26:52 +0000 (12:26 +0100)]
model: support Ministral3 (#17644)
* conversion script
* support ministral 3
* maybe this is better?
* add TODO for rope_yarn_log_mul
* better ppl (tested on 14B-Instruct)
* Add Ministral3 support to Mistral format
* improve arch handling
* add sizes
* Apply suggestions from code review
Co-authored-by: Sigbjørn Skjæret <redacted>
* nits
---------
Co-authored-by: Julien Denize <redacted>
Co-authored-by: Sigbjørn Skjæret <redacted>
Georgi Gerganov [Mon, 1 Dec 2025 10:49:53 +0000 (12:49 +0200)]
metal : add FA head size 48 (#17619)
Georgi Gerganov [Mon, 1 Dec 2025 10:49:33 +0000 (12:49 +0200)]
ggml : extend the GGML_SCHED_NO_REALLOC debug logic of the scheduler (#17617)
Aman Gupta [Mon, 1 Dec 2025 09:12:48 +0000 (17:12 +0800)]
llama-graph: avoid expand_forward for fusion (#17633)
Xuan-Son Nguyen [Sun, 30 Nov 2025 21:51:34 +0000 (22:51 +0100)]
contributing: update guidelines for AI-generated code (#17625)
* contributing: update guidelines for AI-generated code
* revise
Adrien Gallouët [Sun, 30 Nov 2025 21:14:32 +0000 (22:14 +0100)]
cmake : add option to build and link LibreSSL (#17552)
Signed-off-by: Adrien Gallouët <redacted>
Tarek Dakhran [Sun, 30 Nov 2025 20:57:31 +0000 (21:57 +0100)]
model: LFM2-VL fixes (#17577)
* Adjust to pytorch
* Add antialiasing upscale
* Increase number of patches to 1024
* Handle default marker insertion for LFM2
* Switch to flag
* Reformat
* Cuda implementation of antialias kernel
* Change placement in ops.cpp
* consistent float literals
* Pad only for LFM2
* Address PR feedback
* Rollback default marker placement changes
* Fallback to CPU implementation for antialias implementation of upscale
Xuan-Son Nguyen [Sun, 30 Nov 2025 14:33:55 +0000 (15:33 +0100)]
clip: fix nb calculation for qwen3-vl (#17594)
Xuan-Son Nguyen [Sun, 30 Nov 2025 14:32:43 +0000 (15:32 +0100)]
cli: add migration warning (#17620)
Adrien Gallouët [Sun, 30 Nov 2025 12:22:44 +0000 (13:22 +0100)]
common : throttle download progress output to reduce IO flush (#17427)
This change limits progress updates to approximately every 0.1% of the
file size to minimize stdio overhead.
Also fixes compiler warnings regarding __func__ in lambdas.
Signed-off-by: Adrien Gallouët <redacted>
Aaron Teo [Sun, 30 Nov 2025 11:12:32 +0000 (19:12 +0800)]
common: add LLAMA_LOG_FILE env var (#17609)
Signed-off-by: Aaron Teo <redacted>
Gilad S. [Sun, 30 Nov 2025 02:00:59 +0000 (04:00 +0200)]
ggml: fix: macOS build with `-DGGML_BACKEND_DL=ON` (#17581)
ddh0 [Sun, 30 Nov 2025 01:59:25 +0000 (19:59 -0600)]
common: update env var name (#17588)
Aman Gupta [Sun, 30 Nov 2025 00:17:55 +0000 (08:17 +0800)]
CUDA: add stream-based concurrency (#16991)
* CUDA: add stream-based concurrency
* HIP: fix hipStreamWaitEvent define and nodiscard warnings
* ggml-cuda: fix fusion inside stream
* ggml-cuda: fix bug w.r.t first stream launch
* ggml-cuda: format
* ggml-cuda: improve assert message
* ggml-cuda: use lambda instead of duplicating code
* ggml-cuda: add some more comments
* ggml-cuda: add more detailed comments about concurrency
* ggml-cuda: rename + remove unused var
* ggml-cuda: fix condition for stream launch
* ggml-cuda: address review comments, add destructor
* common.cuh: add is_valid for concurrent events
* common.cuh: make comment better
* update comment
Co-authored-by: Johannes Gäßler <redacted>
* update comment
Co-authored-by: Johannes Gäßler <redacted>
* common.cuh: fix lower_bound condition + remove join_node data from write_ranges
* ggml-cuda: fix overlap condition + shadowing parameter
---------
Co-authored-by: Carl Philipp Klemm <redacted>
Co-authored-by: Johannes Gäßler <redacted>
Mahekk Shaikh [Sun, 30 Nov 2025 00:16:28 +0000 (19:16 -0500)]
cuda : add error checking for cudaMemcpyAsync in argsort (#17599)
* cuda : add error checking for cudaMemcpyAsync in argsort (#12836)
* fix indentation
Acly [Sun, 30 Nov 2025 00:03:21 +0000 (01:03 +0100)]
vulkan : fix FA mask load with bounds check (coopmat2) (#17606)
Xuan-Son Nguyen [Sat, 29 Nov 2025 21:04:44 +0000 (22:04 +0100)]
server: move server-context to its own cpp|h (#17595)
* git mv
* add server-context.h
* add server-context.h
* clean up headers
* cont : cleanup
* also expose server_response_reader (to be used by CLI)
* fix windows build
* decouple server_routes and server_http
---------
Co-authored-by: Georgi Gerganov <redacted>
Haiyue Wang [Sat, 29 Nov 2025 17:43:29 +0000 (01:43 +0800)]
server: explicitly set the function name in lambda (#17538)
As [1] explained, the real debug message will be like:
"res operator(): operator() : queue result stop"
Set the name explicitly, the message is easy for debugging:
"res operator(): recv : queue result stop"
The left "operator()" is generated by 'RES_DBG() ... __func__'
[1]: https://clang.llvm.org/extra/clang-tidy/checks/bugprone/lambda-function-name.html
Signed-off-by: Haiyue Wang <redacted>
Igor Smirnov [Sat, 29 Nov 2025 16:06:32 +0000 (21:06 +0500)]
common : fix json schema with '\' in literals (#17307)
* Fix json schema with '\' in literals
* Add "literal string with escapes" test
Neo Zhang [Sat, 29 Nov 2025 12:59:44 +0000 (20:59 +0800)]
sycl : support to malloc memory on device more than 4GB, update the doc and script (#17566)
Co-authored-by: Neo Zhang Jianyu <redacted>
ixgbe [Sat, 29 Nov 2025 12:56:31 +0000 (20:56 +0800)]
ggml: replace hwcap with riscv_hwprobe for RVV detection (#17567)
Signed-off-by: Wang Yang <redacted>
Ruben Ortlam [Sat, 29 Nov 2025 08:37:22 +0000 (09:37 +0100)]
Vulkan: MMVQ Integer Dot K-Quant and MUL_MAT_ID support (#16900)
* vulkan: split mul_mmq_funcs for mul_mat_vecq use
* add mxfp4 mmvq
* add q2_k mmvq
* add q3_k mmvq
* add q4_k and q5_k mmvq
* add q6_k mmvq
* handle 4x4 quants per mmvq thread
* enable MUL_MAT_ID mmvq support
* enable subgroup optimizations for mul_mat_vec_id shaders
* device tuning
* request prealloc_y sync after quantization
* fix indentation
* fix llvmpipe test failures
* fix mul_mat_id mmvq condition
* fix unused variable warning
Jeff Bolz [Sat, 29 Nov 2025 07:39:57 +0000 (01:39 -0600)]
vulkan: improve topk perf for large k, fix overflow in unit tests (#17582)
Aleksei Nikiforov [Fri, 28 Nov 2025 19:53:01 +0000 (20:53 +0100)]
gguf-py : fix passing non-native endian tensors (editor-gui and new-metadata) (#17553)
gguf_new_metadata.py reads data from reader.
Reader doesn't byteswap tensors to native endianness.
But writer does expect tensors in native endianness to convert them
into requested endianness.
There are two ways to fix this: update reader and do conversion to native endianness and back,
or skip converting endianness in writer in this particular USE-case.
gguf_editor_gui.py doesn't allow editing or viewing tensor data.
Let's go with skipping excessive byteswapping.
If eventually capability to view or edit tensor data is added,
tensor data should be instead byteswapped when reading it.
DAN™ [Fri, 28 Nov 2025 18:29:36 +0000 (13:29 -0500)]
common : move all common_chat_parse_* to chat-parser.cpp. (#17481)
o7si [Fri, 28 Nov 2025 18:14:00 +0000 (02:14 +0800)]
server: fix: /metrics endpoint returning JSON-escaped Prometheus format (#17386)
* fix: /metrics endpoint returning JSON-escaped Prometheus format
* mod: remove string overload from ok() method
Diego Devesa [Fri, 28 Nov 2025 15:33:23 +0000 (07:33 -0800)]
ggml : add GGML_SCHED_NO_REALLOC option to disable reallocations in ggml_backend_sched (#17276)
* ggml : add GGML_SCHED_NO_REALLOC option to disable reallocations in ggml_backend_sched
Enabled in ggml-ci for testing.
* llama : update worst-case graph for unified cache
* ci : disable op offload in some tests
* fix spelling
---------
Co-authored-by: Georgi Gerganov <redacted>
R0CKSTAR [Fri, 28 Nov 2025 13:08:29 +0000 (21:08 +0800)]
[MUSA] enable fp16/fast_fp16/bf16_mma on PH1 (#17551)
* [MUSA] enable fp16/fast_fp16/bf16_mma on PH1
Signed-off-by: Xiaodong Ye <redacted>
* Update ggml/src/ggml-cuda/fattn-vec.cuh
Co-authored-by: Johannes Gäßler <redacted>
* Update ggml/src/ggml-cuda/fattn-vec.cuh
Co-authored-by: Johannes Gäßler <redacted>
* Update ggml/src/ggml-cuda/fattn-tile.cuh
Co-authored-by: Johannes Gäßler <redacted>
* Address review comments
Signed-off-by: Xiaodong Ye <redacted>
---------
Signed-off-by: Xiaodong Ye <redacted>
Co-authored-by: Johannes Gäßler <redacted>
Aman Gupta [Fri, 28 Nov 2025 12:34:51 +0000 (20:34 +0800)]
ggml-cuda: add stricter checking for fusion (#17568)
* ggml-cuda: make conditions for fusion more explicit
* ggml-cuda: remove size check as std::equal already does it
Fredrik Hultin [Fri, 28 Nov 2025 11:57:04 +0000 (12:57 +0100)]
server : add Anthropic Messages API support (#17570)
* server : add Anthropic Messages API support
* remove -@pytest.mark.slow from tool calling/jinja tests
* server : remove unused code and slow/skip on test_anthropic_vision_base64_with_multimodal_model in test_anthropic_api.py
* server : removed redundant n field logic in anthropic_params_from_json
* server : use single error object instead of error_array in streaming response handler for /v1/chat/completions and use unordered_set instead of set in to_json_anthropic_stream()
* server : refactor Anthropic API to use OAI conversion
* make sure basic test always go first
* clean up
* clean up api key check, add test
---------
Co-authored-by: Xuan Son Nguyen <redacted>
Piotr Wilkin (ilintar) [Fri, 28 Nov 2025 11:02:56 +0000 (12:02 +0100)]
model : Qwen3 Next (#16095)
* Qwen3 Next - cleaned up version
* Whitespaces and stuff
* Correct minor errors
* Update src/llama-model.cpp
Co-authored-by: Sigbjørn Skjæret <redacted>
* Misc. fixes.
* Clean up code, add missing hybrid qualifier
* Did someone transpose the SOLVE_TRI result matrix? Perhaps...
* Whitespace
* Proper tensors for cb calls
* Use llama-graph.h vertical alignment
* BROKEN: chunking
* Set new tensors as inputs.
* Proper chunk logic
* It's the circle of life...
* More shenanigans for n_seq > 1
* Nail in the coffin?
* Fix Windows build
* Eh, one fails on Windows, the other fails on Mac... just use general capture.
* quant : cleanup
* model : cleanup
* qwen3 : cleanup
* cont : cleanup
* cont : cleanup
* ggml : revert change
* qwen3 : cleanup
* cont : cleanup
* Readd cmath
* qwen3 : fix typo
* Update convert_hf_to_gguf.py
Co-authored-by: Sigbjørn Skjæret <redacted>
* Usual suspects
* fix my bad suggestion
---------
Co-authored-by: Sigbjørn Skjæret <redacted>
Co-authored-by: Georgi Gerganov <redacted>
Johannes Gäßler [Fri, 28 Nov 2025 09:29:09 +0000 (10:29 +0100)]
CUDA: no FP16 arithmetic for vector FA kernel (#17558)
Jeff Bolz [Fri, 28 Nov 2025 09:07:29 +0000 (03:07 -0600)]
vulkan: Implement GGML_OP_TRI (#17503)
* vulkan: Implement GGML_OP_TRI
* check types match
Radoslav Gerganov [Fri, 28 Nov 2025 08:33:51 +0000 (10:33 +0200)]
rpc : cache and reuse compute graphs (#15405)
Store the last computed graph and reuse it when possible.
Also do not return response from GRAPH_COMPUTE and assume it always
completes successfully. If this this is not the case, the server closes
the connection. This saves us a network round trip to the server.
yulo [Fri, 28 Nov 2025 07:24:30 +0000 (15:24 +0800)]
HIP: enable mul_mat_f for RDNA4 (#17437)
* enable mmf for rdna4
* move some mmvf to mmf
* revert lds128 for wmma loading
* Revert "revert lds128 for wmma loading"
This reverts commit
db9ae8b6b4738a5def5b393caa1611d52133e9b5 .
* Revert "enable mmf for rdna4"
This reverts commit
698c9f24187b990e35c3b73a8067e5387e6ddbd4 .
* Revert "move some mmvf to mmf"
This reverts commit
99b92bd6653cc8593607f641e44606391691792f .
* enable mul_mat for rdna4
---------
Co-authored-by: zhang hui <redacted>
Piotr Wilkin (ilintar) [Fri, 28 Nov 2025 04:15:32 +0000 (05:15 +0100)]
SOLVE_TRI CUDA kernel for small matrices (#17457)
Neo Zhang Jianyu [Fri, 28 Nov 2025 00:50:56 +0000 (08:50 +0800)]
refactor pad_reflect_1d to make the UT case pass (#17204)
Co-authored-by: Zhang Jianyu <redacted>
Jeff Bolz [Thu, 27 Nov 2025 14:48:00 +0000 (08:48 -0600)]
vulkan: Implement SOLVE_TRI (#17486)
* vulkan: Implement SOLVE_TRI
* load B matrix through shared memory
* use FLOAT_TYPE
Georgi Gerganov [Thu, 27 Nov 2025 14:34:13 +0000 (16:34 +0200)]
arch : add description about LLM_TENSOR_INFOS (#17550)
Georgi Gerganov [Thu, 27 Nov 2025 14:04:29 +0000 (16:04 +0200)]
models : fix LFM2 tensors (#17548)
matt23654 [Thu, 27 Nov 2025 11:35:35 +0000 (11:35 +0000)]
cuda : fix UMA detection on discrete GPUs. (#17537)
Alberto Cabrera Pérez [Thu, 27 Nov 2025 11:25:14 +0000 (11:25 +0000)]
ggml-cpu: aarm64: q4_K repack gemm and gemv implementations (dotprod only) (#17494)
* Enabled q4_K_4x8 path
* Fixed generic Q4_K 8x4 implementation
* wip: dotprod gemm
* Working arm q4_K dotprod gemm
Signed-off-by: Alberto Cabrera <redacted>
* Undo acc rename
Signed-off-by: Alberto Cabrera <redacted>
* Q4_K arm dotprod gemm
Signed-off-by: Alberto Cabrera <redacted>
* Fix: q4_qs reinterpret from uint to int
Signed-off-by: Alberto Cabrera <redacted>
* Removed comments
* Fixed macro guards
* Fixed unused vars in generic implementation
* Fixed unused vars in 8x4 repack
* Fixed unused vars in generic implementation, unneeded comment
* Missing arch fallback for x86
* minor : style
---------
Signed-off-by: Alberto Cabrera <redacted>
Co-authored-by: Georgi Gerganov <redacted>
Eric Curtin [Thu, 27 Nov 2025 10:35:47 +0000 (10:35 +0000)]
devops: Add build-essential to Ubuntu 26.04 image (#17531)
This is no longer passing the build, needs more packages.
Signed-off-by: Eric Curtin <redacted>
Aleksei Nikiforov [Thu, 27 Nov 2025 10:35:38 +0000 (11:35 +0100)]
gguf-py : skip endian-conversion of MXFP4 data (#17523)
* gguf_convert_endian.py: skip MXFP4 data
* Use gguf.constants.GGML_QUANT_SIZES to determine block sizes
Acly [Thu, 27 Nov 2025 05:54:19 +0000 (06:54 +0100)]
vulkan : move contiguous checks to device_supports_op (#17490)
* vulkan : remove op_supports_incontiguous and add missing constraints in device_supports_op
* im2col: remove contraints on src0 (kernel input)
Jeff Bolz [Thu, 27 Nov 2025 05:32:30 +0000 (23:32 -0600)]
vulkan: use a fixed 1KB buffer for the add_rms_fusion opt (#17514)
Xuan-Son Nguyen [Thu, 27 Nov 2025 00:02:50 +0000 (01:02 +0100)]
server: enable jinja by default, update docs (#17524)
* server: enable jinja by default, update docs
* fix tests
lhez [Wed, 26 Nov 2025 21:29:58 +0000 (13:29 -0800)]
opencl: add sqr, sqrt, mean and ssm_conv (#17476)
* opencl: add sqr
* opencl: add sqrt
* opencl: add mean
* opencl: add ssm_conv
* opencl: add missing cl_khr_fp16
* opencl: do sqrt in f32 then convert to f16 for better precision
Alberto Cabrera Pérez [Wed, 26 Nov 2025 21:14:54 +0000 (21:14 +0000)]
Fix chunks being too small with small matrix sizes (#17526)
Han Qingzhe [Wed, 26 Nov 2025 20:44:07 +0000 (04:44 +0800)]
clip: (minicpmv) fix resampler kq_scale (#17516)
* debug:"solve minicpmv precision problem"
* “debug minicpmv”
* Apply suggestion from @ngxson
---------
Co-authored-by: Xuan-Son Nguyen <redacted>
Jeff Bolz [Wed, 26 Nov 2025 15:46:33 +0000 (09:46 -0600)]
vulkan: allow graph_optimize for prompt processing workloads (#17475)
Jeff Bolz [Wed, 26 Nov 2025 15:45:43 +0000 (09:45 -0600)]
vulkan: Implement top-k (#17418)
* vulkan: Implement top-k
Each pass launches workgroups that each sort 2^N elements (where N is usually 7-10)
and discards all but the top K. Repeat until only K are left. And there's a fast
path when K==1 to just find the max value rather than sorting.
* fix pipeline selection
* vulkan: Add N-ary search algorithm for topk
* microoptimizations
xctan [Wed, 26 Nov 2025 13:33:05 +0000 (21:33 +0800)]
ggml-cpu : add RISC-V Zvfh impl for ggml_vec_mad_f16 (#17448)
* ggml-cpu : add RISC-V Zvfh impl for ggml_vec_mad_f16
* ggml-cpu : dedup scalar impl
* Update ggml/src/ggml-cpu/vec.h
---------
Co-authored-by: Georgi Gerganov <redacted>
Adrien Gallouët [Wed, 26 Nov 2025 13:15:21 +0000 (14:15 +0100)]
cmake : use EXCLUDE_FROM_ALL to avoid patch-boringssl.cmake (#17520)
We have to separate the code path starting 3.28 because
`FetchContent_Populate` is now deprecated and will be completely removed
in a future version.
Signed-off-by: Adrien Gallouët <redacted>
Adrien Gallouët [Wed, 26 Nov 2025 13:14:41 +0000 (14:14 +0100)]
ggml : fix ARM feature verification (#17519)
On arm64 with `cmake` version 3.31.6, the final feature verification fails:
-- ARM detected flags: -mcpu=neoverse-v2+crc+sve2-aes+sve2-sha3+nossbs
-- Performing Test GGML_MACHINE_SUPPORTS_dotprod
-- Performing Test GGML_MACHINE_SUPPORTS_dotprod - Success
-- Performing Test GGML_MACHINE_SUPPORTS_i8mm
-- Performing Test GGML_MACHINE_SUPPORTS_i8mm - Success
-- Performing Test GGML_MACHINE_SUPPORTS_sve
-- Performing Test GGML_MACHINE_SUPPORTS_sve - Success
-- Performing Test GGML_MACHINE_SUPPORTS_sme
-- Performing Test GGML_MACHINE_SUPPORTS_sme - Failed
-- Performing Test GGML_MACHINE_SUPPORTS_nosme
-- Performing Test GGML_MACHINE_SUPPORTS_nosme - Success
-- Checking for ARM features using flags:
-- -U__ARM_FEATURE_SME
-- -mcpu=neoverse-v2+crc+sve2-aes+sve2-sha3+nossbs+dotprod+i8mm+sve+nosme
-- Performing Test HAVE_DOTPROD
-- Performing Test HAVE_DOTPROD - Failed
-- Performing Test HAVE_SVE
-- Performing Test HAVE_SVE - Failed
-- Performing Test HAVE_MATMUL_INT8
-- Performing Test HAVE_MATMUL_INT8 - Failed
-- Performing Test HAVE_FMA
-- Performing Test HAVE_FMA - Success
-- Performing Test HAVE_FP16_VECTOR_ARITHMETIC
-- Performing Test HAVE_FP16_VECTOR_ARITHMETIC - Failed
-- Performing Test HAVE_SME
-- Performing Test HAVE_SME - Failed
-- Adding CPU backend variant ggml-cpu: -U__ARM_FEATURE_SME;-mcpu=neoverse-v2+crc+sve2-aes+sve2-sha3+nossbs+dotprod+i8mm+sve+nosme
We need to explicitly replace `;` with spaces from the list to make
`CMAKE_REQUIRED_FLAGS` work correctly...
Signed-off-by: Adrien Gallouët <redacted>
Jiacheng (Jason) Chen [Wed, 26 Nov 2025 10:18:48 +0000 (05:18 -0500)]
HIP: Patch failed testcase in WMMA-MMQ kernels for RDNA 4 (#17502)
* patch failed test case MUL_MAT(type_a=q4_0,type_b=f32,m=576,n=512,k=576,bs=[1,1],nr=[1,1],per=[0,1,2,3],k_v=0,o=1) for enabling WMMA on RDNA4
* Quick clean up on mma.cuh to add ggml_cuda_memcpy_1 back in for half2 and bfloat162
hipudding [Wed, 26 Nov 2025 08:44:19 +0000 (16:44 +0800)]
CANN: Add MROPE and IMROPE support (#17401)
* CANN: ROPE supports both MROPE and IMROPE.
1. Optimize the caching logic of rope_cache_init.
2. Add support for mRoPE and i-mRoPE.
Note that on Ascend 910B devices, it is necessary to disable FA
in CLIP and disable NZ-format conversion. These two issues are
still under investigation.
* Resolve review comments
o7si [Wed, 26 Nov 2025 07:21:06 +0000 (15:21 +0800)]
chore: upgrade cpp-httplib from v0.27.0 to v0.28.0 (#17513)
Jeff Bolz [Wed, 26 Nov 2025 06:08:10 +0000 (00:08 -0600)]
vulkan: Implement GGML_OP_CUMSUM (#17479)
Georgi Gerganov [Tue, 25 Nov 2025 13:31:43 +0000 (15:31 +0200)]
ggml : add ggml_top_k (#17365)
* ggml : add ggml_top_k
* cont : add ggml_argsort_top_k
* metal : add top_k support
* ggml : cleanup
* tests : add virtual err() function for test_case
* ggml : add comments
Aleksei Nikiforov [Tue, 25 Nov 2025 13:18:16 +0000 (14:18 +0100)]
convert : fix big-endian conversion (#17431)
* Fix convert_hf_to_gguf.py script on s390x
Assume converted model data is originally little-endian.
Byteswap data on s390x after reading it to put values in correct presentation
for any transformation needed, like calculating weight tensors.
Then byteswap data to little-endian before passing it to GGUFWriter while
GGUFWriter will byteswap data back to big endian if big endian output is requested.
byteswap(inplace=True) calls don't work with lazy tensor and array wrappers.
Use byteswap with copying data to workaround this behaviour.
* Make GGUFWriter accept tensors in native endianness instead of little-endian
With this change if no byteswapping is actually needed, 2 excessive byteswaps can be omitted on s390x
* Fix byteswapping in convert_hf_to_gguf.py for remote models
Diego Devesa [Tue, 25 Nov 2025 12:00:23 +0000 (04:00 -0800)]
codeowners : remove slaren (#17492)
TianHao324 [Tue, 25 Nov 2025 09:39:06 +0000 (17:39 +0800)]
CANN: supports out_prod operator for F32 and F16 (#17406)
Co-authored-by: tianhao <redacted>
Pascal [Tue, 25 Nov 2025 07:01:02 +0000 (08:01 +0100)]
webui: add rehype plugin to restore HTML in Markdown table cells (#17477)
* webui: add rehype plugin to restore HTML in Markdown table cells
The remark/rehype pipeline neutralizes inline HTML as literal text
(remarkLiteralHtml) so that XML/HTML snippets in LLM responses display
as-is instead of being rendered. This causes <br> and <ul> markup in
table cells to show as plain text.
This plugin traverses the HAST post-conversion, parses whitelisted HTML
patterns (<br>, <ul><li>) from text nodes, and replaces them with actual
HAST element nodes. For lists, adjacent siblings must be combined first
as the AST fragmentation breaks pattern matching.
Strict validation rejects malformed markup, keeping it as raw text.
* chore: update webui build output
Jeff Bolz [Tue, 25 Nov 2025 06:11:27 +0000 (00:11 -0600)]
vulkan: Use fewer rows for scalar FA when HS is not a multiple of 16 (#17455)
Aaron Teo [Tue, 25 Nov 2025 01:56:07 +0000 (09:56 +0800)]
llama: introduce support for model-embedded sampling parameters (#17120)
Jeff Bolz [Mon, 24 Nov 2025 21:25:24 +0000 (15:25 -0600)]
vulkan: more FA details in vk_perf_logger (#17443)
Daniel Bevenius [Mon, 24 Nov 2025 20:06:17 +0000 (21:06 +0100)]
llama : skip output reordering for single token batches (#17466)
This commit adds a check to skip the output reordering logic when
n_outputs == 1. With a single output token, the data is trivially
sorted and the reordering code is currently doing unnecessary work
(resetting and rebuilding output_ids to the same values).
The motivation for this change is improved code clarity and avoiding
confusion when debugging. While the performance impact is probably
negligible, this unnecessary work happens on every decode call in
llama-server when processing batches with single-token outputs.
Jiacheng (Jason) Chen [Mon, 24 Nov 2025 19:00:10 +0000 (14:00 -0500)]
HIP: WMMA-MMQ kernels for RDNA 4 (#17156)
* first commit naive test to enable mmq for RDNA4
* adding appropriate WMMA instructions
* git rebase on top of master: fixing the correctness of the mat mul operations, updating layout mappings for RDNA4
* clean up merge conflicts
* add comments and code clean up
* PR clean up, addressed comments
* enable MMQ fallback on RDNA4
* addressed comments: add guards in load generic, separate wmma branch for use_mmq function
* Revert build-xcframework.sh
* Formating: remove trailing whitespace
* revert CMake files
* clean up after rebase: remove duplicated change, revert cmake files
* clean up after rebase: revert changes from build-xcframework.sh
* clean up: remove extra space line in mma.cuh
* Revert "clean up: remove extra space line in mma.cuh"
This reverts commit
b39ed57c4529906466bd0bc7c2a86e08fc2f8bee .
Sigbjørn Skjæret [Mon, 24 Nov 2025 14:50:55 +0000 (15:50 +0100)]
convert : allow quantizing lora again (#17453)
Xuan-Son Nguyen [Mon, 24 Nov 2025 13:41:53 +0000 (14:41 +0100)]
server: split server.cpp code into server/common/task/queue (#17362)
* add server-task, server-common
* add server-queue
* rm redundant includes
* move enum stop_type to server-task
* server : headers cleanup
---------
Co-authored-by: Georgi Gerganov <redacted>
Daniel Bevenius [Mon, 24 Nov 2025 13:38:45 +0000 (14:38 +0100)]
examples : add -kvu to batched usage example [no ci] (#17469)
This commit adds the --kv-unified flag to the usage example
in the README.md file for the batched example.
The motivation for this is that without this flag the example will fail
with the following error:
```console
Hello my name is
split_equal: sequential split is not supported when there are coupled
sequences in the input batch (you may need to use the -kvu flag)
decode: failed to find a memory slot for batch of size 4
main: llama_decode() failed
```
Georgi Gerganov [Mon, 24 Nov 2025 12:28:37 +0000 (14:28 +0200)]
sync : ggml
Daniel Bevenius [Mon, 24 Nov 2025 11:51:50 +0000 (12:51 +0100)]
ggml : remove dirty flag from version string (ggml/1391)
This commit removes the "-dirty" suffix from the GGML version string.
The motivation for this change is to ensure that the version string
works with different ways of checking out ggml and using it in projects.
By removing the dirty flag from the version string, we avoid potential
artifacts like shared libraries getting a -dirty suffix in their names.
Instead, if the project is built from a dirty git state, the dirty flag
will be appended to the commit hash in the GGML_BUILD_COMMIT variable.
This will enable users to still identify that the build was made from
from a modified/dirty state even though the version might match a "real"
version.
For example, the commit can be produces as follows:
```c++
printf("commit: %s\n", ggml_commit());
```
Which would print the following for a dirty build:
```console
commit:
781baf2a -dirty
```
Refs: https://github.com/ggml-org/ggml/pull/1363#issuecomment-
3569691546
Alberto Cabrera Pérez [Mon, 24 Nov 2025 11:08:11 +0000 (11:08 +0000)]
ggml-cpu: arm64: q4_K repack gemm and gemv implementations (i8mm) (#16739)
* Enabled q4_K_8x8_q8_K path on ARM
* wip: I8mm qs multiplication, pending bias
* cpu : arm : REPACK gemm q4_K8x8 implementation
Signed-off-by: Alberto Cabrera <redacted>
* Guard gemm with proper features, improved superblock scale and min calc
Signed-off-by: Alberto Cabrera <redacted>
* cpu: arm: Implemented REPACK gemv for Q4_K
Signed-off-by: Alberto Cabrera <redacted>
* Removed completed TODO
* Fixed missing guards when selecting optimal repack type for Q4_K
Signed-off-by: Alberto Cabrera <redacted>
* Fixed macro guard for gemv
* Fixed wrong comment in GEMV
* Fixed warning for unused variable
* vdotq_s32 -> ggml_vdotq_s32
Signed-off-by: Alberto Cabrera <redacted>
* Clang-format issues
* Apply suggestions from code review
Co-authored-by: Diego Devesa <redacted>
* Removed unnecessary GGML_UNUSED
* Fixed guards in q4_k gemm and gemv (repack)
---------
Signed-off-by: Alberto Cabrera <redacted>
Co-authored-by: Diego Devesa <redacted>
ixgbe [Mon, 24 Nov 2025 11:07:14 +0000 (19:07 +0800)]
ggml: add RISC-V cpu-feats (#17461)
* ggml: add RISC-V cpu-feats
Signed-off-by: Wang Yang <redacted>
* fix comment[1]
---------
Signed-off-by: Wang Yang <redacted>
william pan [Mon, 24 Nov 2025 06:16:56 +0000 (22:16 -0800)]
models : Added support for RND1 Diffusion Language Model (#17433)
* Converted RND1 model to GGUF weights
* RND1 llama.cpp support v1
* RND1 llama.cpp support v2 non causal bug
* RND1 llama.cpp support v3 doccumentation
* RND1 llama.cpp support v4 clean code
* linting issues
* RND1 pr fixes v1
* RND1 pr fixes v2
Co-authored-by: Sigbjørn Skjæret <redacted>
* Diffusion documentation edits
---------
Co-authored-by: Sigbjørn Skjæret <redacted>
Max Krasnyansky [Mon, 24 Nov 2025 02:55:56 +0000 (18:55 -0800)]
hexagon: add support for ROPE_NEOX (#17458)
Raul Torres [Mon, 24 Nov 2025 02:02:52 +0000 (02:02 +0000)]
CANN: Define `cann_graph_update_required` before macro (#17434)
**Description of the problem**
`cann_graph_update_required` is redundantly defined and
initialized as `false` inside two mutually exclusive macro branches.
**Proposed solution**
Define it right before the macro so that it could serve both
branches.
M. Mediouni [Mon, 24 Nov 2025 00:54:49 +0000 (01:54 +0100)]
ggml-hexagon: Initial Hexagon v68/v69 support (#17394)
* ggml-hexagon: fix build error with GCC
Add stdexcept include to fix GCC build errors
Signed-off-by: Mohamed Mediouni <redacted>
* ggml-hexagon: check VTCM acquire failures
Signed-off-by: Mohamed Mediouni <redacted>
* ggml-hexagon: disable destination bypass on older than v73
v68 errors out if having bypass enabled when the VTCM is the destination.
At least on v68 this made things actually work... not a proper fix though, so to look at later...
Signed-off-by: Mohamed Mediouni <redacted>
* ggml-hexagon: add initial v68/v69 support
v68 is the Hexagon revision notably used on the Snapdragon 8cx
Gen 3 and the QCM6490.
Also add support for v69.
8MB isn't a supported page size, so relax asked for page size constraint
for HAP_compute_res_attr_set_vtcm_param_v2 to optimal.
Signed-off-by: Mohamed Mediouni <redacted>
---------
Signed-off-by: Mohamed Mediouni <redacted>