12a4a47e6a
Fix GLM 4.7 Lite MoE gating func ( #18980 )
...
* Fix GLM 4.7 MoE gating func
* Update src/models/deepseek2.cpp
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
* Update src/llama-model.cpp
Co-authored-by: Xuan-Son Nguyen <thichthat@gmail.com>
---------
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
Co-authored-by: Xuan-Son Nguyen <thichthat@gmail.com>
2026-01-21 12:35:20 +01:00
37c35f0e1c
gguf: display strerrno when cant load a model ( #18884 )
...
I've had issues loading models with llama-server:
[44039] E gguf_init_from_file: failed to open GGUF file 'mistral-7b-v0.1.Q8_0.gguf'
and I was sure it could access the file. Seems like --models-dir and
--models-presets dont interact like I thought they would but I salvaged
this snippet that helps troubleshooting
[44039] E gguf_init_from_file: failed to open GGUF file 'mistral-7b-v0.1.Q8_0.gguf' (errno No such file or directory)
2026-01-21 08:52:46 +02:00
5bd341c9a1
CUDA: Fix builds for older CCCL versions by ifdefing strided_iterator ( #18964 )
...
* CUDA: Fix builds for older CCCL versions by ifdefing strided_iterator
Strided iterator was added in [CCCL
3.1](https://github.com/NVIDIA/cccl/releases/tag/v3.1.0 ), which is packaged into
[CTK
13.1](https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html#id5 )
* Unindent as per code review request
2026-01-21 02:34:29 +01:00
1c7cf94b22
common, server : use the same User-Agent by default ( #18957 )
...
This commit also ensures that if a custom User-Agent is used, it will be
the only one sent.
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-01-20 18:28:43 +01:00
2c1f199653
cli : fix reasoning responses in CLI ( #18961 )
...
* cli : fix reasoning responses in CLI
* fix build
* fix build (2)
2026-01-20 18:23:25 +01:00
d1e3556481
CUDA: Replace init_offsets kernel with iterators in cub-based argsort ( #18930 )
...
* CUDA: Replace `init_offsets` with iterators in argsort
This is a QOL improvement, saving us the cost of materializing the
iterator
* Remove unnecessary include from top-k.cu
2026-01-20 20:11:01 +08:00
08f3f4a8a3
ggml : cleanup path_str() ( #18928 )
...
- Remove pragmas as `std::codecvt_utf8` is not used.
- Avoid implicit `strlen()`.
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-01-20 11:42:49 +01:00
271191906c
metal : enable FA for MLA heads ( #18950 )
2026-01-20 12:21:28 +02:00
7dee9ff59a
convert : use n_groups instead of hardcoded values in reshape ( #18929 )
...
* convert : use n_groups instead of hardcoded values in reshape
This commit modifies the conversion script for NemotronHModel to use
the 'n_groups' hyperparameter, and allow Python to calculate the the
last dimension, using -1, when reshaping the 'mixer.norm.weight' tensor.
* use self.n_group instead of self.hparams["n_groups"]
2026-01-20 06:55:24 +01:00
6df686bee6
server : refactor oai_parser_opt, move it to server_chat_params ( #18937 )
...
* server_chat_params
* move chat format into CLI
* use meta whenever possible
* clean up, no more chatml fallback
2026-01-19 23:28:01 +01:00
1706a6d7c6
convert : support Glm4MoeLite ( #18936 )
...
* initial commit for branch
* add glm-4.7-flash, move tokenizer hash
* use `glm4` pretok
* silence flake8 E302 (CI)
* apply review feedback
* add <|user|> as eog
* also add EOG `<|observation|>`
* revert llama-vocab
* inherit vocab from glm4
---------
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
2026-01-19 23:09:20 +01:00
959ecf7f23
jinja : fix undefined keys and attributes and int/float as bool ( #18924 )
...
* fix undefined keys and attributes
* add falsy tests
* as_bool for integers and floats
* more falsy/truthy tests
* --typo
2026-01-19 20:29:43 +01:00
4037093c66
ci : run test-jinja -py on high perf [no ci] ( #18916 )
2026-01-19 20:29:15 +01:00
18361c579c
server: fix memory reservations in populate_token_probs ( #18787 )
2026-01-19 19:13:31 +01:00
365a3e8c31
ggml : add ggml_build_forward_select ( #18550 )
...
* ggml : add ggml_build_forward_select
* cuda : adapt CUDA graph compat to new feature
* vulkan : update logic to handle command buffer closing
* ggml : check compute for fusion
* ggml : add comment
2026-01-19 20:03:19 +02:00
39d0ff485d
chore: update webui build output
2026-01-19 19:02:40 +01:00
8a95ec3ea6
feat: Improve MCP Server selection UI + lazy load health checks
2026-01-19 19:01:32 +01:00
cafb9c09d3
feat: UI improvements
2026-01-19 16:56:02 +01:00
54192b05fb
feat: Simplify MCP server enabling logic per chat
...
Refactors MCP server enabling logic to remove the dependency on global settings.
This simplifies the logic by directly checking the per-chat override status, and removes the need to pass the global enabled state as a parameter.
Additionally:
- Only shows MCP servers that are enabled in settings in the selector.
- Sorts the servers by whether they are enabled for the current chat.
2026-01-19 16:43:53 +01:00
62ed7f112d
chore: update webui build output
2026-01-19 16:26:16 +01:00
d37683942b
fix: Missing onModelChange callback running assistant message re-generation
2026-01-19 16:25:49 +01:00
3d55846a5c
model-conversion : add BUILD_DIR variable to run-converted-model scripts ( #18927 )
...
This commit adds a BUILD_DIR variable to the scripts used for running
converted models.
The motivation for this is that currently the `build` directory is
hardcoded and it can be useful to specify a different build directory,
with builds for different configurations.
2026-01-19 13:12:38 +01:00
d6dfe8e064
chore: update webui build output
2026-01-19 12:12:52 +01:00
058929d453
fix: acurate tool_response display
2026-01-19 12:11:06 +01:00
287a33017b
llama : Extend fallback, fix fileno for dio file, exclude case that mmap uses dio file ( #18887 )
2026-01-18 18:35:57 +02:00
d92b621346
fix: unify MCP server label logic with simplified fallback
2026-01-18 13:10:03 +01:00
293a1565dc
docs: add linux to index ( #18907 )
2026-01-18 18:03:35 +08:00
16a03eea36
chore: update webui build output
2026-01-18 10:43:45 +01:00
d8af98f1ed
refactor: remove multimodal validation from model selector
...
Remove all frontend validation logic that prevented users from selecting
models based on multimodal capabilities. This refactoring removes
restrictive UI code while maintaining full functionality
- Vision models can describe images as text
- That text remains useful for non-vision models
- Chaining vision -> non-vision is a valid workflow
- Users know their use case better than the UI
- Users can return to vision models when needed
2026-01-18 10:42:01 +01:00
fe44d35574
tests : add test-jinja -py option for cross-checking ( #18906 )
...
* tests : add test-jinja -py option or cross-checking
* Update tests/test-jinja.cpp
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
* fix + add source
* SandboxedEnvironment
* fix array.map case
---------
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-01-18 08:14:27 +01:00
bbcdac0189
jinja : fix object item order (and properly implement dictsort) ( #18904 )
...
* fix object item order
* as_ordered_object
* copy whole object
2026-01-18 03:40:06 +01:00
d03c45c9c5
jinja : attribute support for join, map and sort ( #18883 )
...
* support negative array index and default value
* attribute support (int and str) for join, map and sort
* add tests
* update CODEOWNERS
* improve fixme sorting comment
2026-01-18 02:53:01 +01:00
10c98cbdf6
jinja : add missing tojson filter for bool ( #18900 )
...
* add missing tojson for bool
* add more literal tests
2026-01-18 01:05:09 +01:00
420960ab92
jinja : fix lexing of float literals with sign ( #18901 )
...
* fix lexing of float literals with sign
* add test
* consume_numeric
2026-01-18 00:57:51 +01:00
f55b033ae6
jinja: correct member access rule ( #18905 )
2026-01-18 00:48:55 +01:00
d1b4757ded
opencl: fix q6_K mv for m=1 ( #18893 )
2026-01-17 13:50:32 -08:00
57c0beaed0
ci : add label for jinja changes ( #18903 )
2026-01-17 21:52:02 +01:00
5c28b7a2ee
chore: update webui build output
2026-01-17 18:38:50 +01:00
fca7177eae
fix: ignore assistant attachments (MCP) for modality detection
2026-01-17 18:36:41 +01:00
3572667788
chore: update webui build output
2026-01-17 16:35:54 +01:00
506da17931
refactor: eliminate MCP circular dependency
...
- Change architecture from mcpStore <-> mcpClient to mcpClient -> mcpStore
- Remove bidirectional callback pattern (set*Callback, notify* methods)
- Add updateState/updateHealthCheck public methods in mcpStore
- Replace callback calls with direct mcpStore method calls
- Remove unused imports (browser, HealthCheckState) and constructor
- Fixes CI: ReferenceError Cannot access mcpClient before initialization
2026-01-17 16:30:42 +01:00
9b3417703f
fix: remove obsolete modality UI tests causing CI failures
...
- Remove VisionModality/AudioModality test stories
- Remove mockServerProps usage and imports
- Simplify Default test (remove dropdown interaction checks)
- Simplify FileAttachments test (remove mocks)
2026-01-17 16:30:36 +01:00
2fbde785bc
kv-cache : optimize KQ mask construction ( #18842 )
...
* kv-cache : optimize KQ mask construction
* cont : add explanation + improve
* cont : fix
2026-01-17 15:42:42 +02:00
a89002f07b
ggml webgpu: support for backend sampling ( #18880 )
...
* ggml webgpu: add SOFTPLUS unary operator
Implements SOFTPLUS (log(1 + exp(x))) with f16/f32 support. Uses f32
precision for intermediate calculations to prevent f16 overflow.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* Follow Vulkan backend numerical stability pattern
* ggml webgpu: add EXPM1 unary operator
Implements EXPM1 (exp(x) - 1) with f16/f32 support.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* ggml webgpu: add FLOOR unary operator
Implements FLOOR (rounds down to nearest integer) with f16/f32 support.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* ggml webgpu: add CEIL unary operator
Implements CEIL (rounds up to nearest integer) with f16/f32 support.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* ggml webgpu: add ROUND unary operator
Implements ROUND (rounds to nearest integer) with f16/f32 support.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* ggml webgpu: add TRUNC unary operator
Implements TRUNC (truncates towards zero) with f16/f32 support.
* Add shader implementation and 4 variants (f32/f16, inplace/non-inplace)
* Register pipelines and device support
* docs : update WebGPU support for unary operators (FLOOR, CEIL, ROUND, TRUNC, EXPM1, SOFTPLUS)
* Updates to webgpu get_memory
* Add argmax
* Add argmax,cumsum,sum,sum_rows
* Add necessary CPY/GET_ROWS operators
* Support for argsort using multi-pass strategy
* Update set_rows for i32 indices, move to pre-wgsl
* Port unary operators to pre-wgsl and support FILL
* Implement PAD
* Add support for top-k
* clean up, scope pipeline init mutex
* fix newline
* Add support for log
* Update LOG for better precision, and ops doc
---------
Co-authored-by: Abhijit Ramesh <abhijitramesh2k@gmail.com>
2026-01-16 16:12:43 -08:00
a723238245
chore: update webui build output
2026-01-16 19:52:23 +01:00
229aba7c3e
fix: strip reasoning content and UI proprietary tags from prompts
...
TODO: add toggle and ensure backend API compliance for reasoning format
2026-01-16 19:50:36 +01:00
388ce82241
ggml : extend ggml_pool_1d + metal ( #16429 )
...
* chore: resolve conflicts
* feat: ggml metal impl
* fix: ggml_metal_kargs_pool_1d struct
* fix: require contiguous input
* chore: test pool_1d
* chore: limit pool1d test cases to p0=0 and s0=k0 to conform with asserts
* chore: add p0 and s0 to testing
* fix: allow padding for cpu and metal
* Update ggml/src/ggml-metal/ggml-metal.metal
* fix: correct single-threaded loop
* ggml : cleanup
* tests : add ne[1] != 1 tests
* fix: ne[1] handling in np
* cont : fixes
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-01-16 16:59:56 +02:00
f09395821b
chore: update webui build output
2026-01-16 15:22:46 +01:00
78c6380222
refactor: remove reasoning after first turn filter
2026-01-16 15:19:50 +01:00
2973c64609
refactor: inline reasoning with tags, remove fixed thinking field
2026-01-16 15:19:42 +01:00
Piotr Wilkin (ilintar)
Matthieu Coudron
Oliver Simons
Adrien Gallouët
Xuan-Son Nguyen
Georgi Gerganov
Daniel Bevenius
ddh0
Sigbjørn Skjæret
Lennart Austenfeld
Aleksander Grygier
Pascal
Julius Tischbein
Francisco Herrera
lhez
Reese Levine
Thore Koritzius