Commit graph

9105 commits

Author SHA1 Message Date
Vitaly Chikunov
e1130f67db ALT: Re-generate tools/server/public/index.html.gz
+ mkdir /usr/src/.npm-global
+ npm config set prefix /usr/src/.npm-global
+ npm install -g @aikidosec/safe-chain

added 57 packages in 6s

5 packages are looking for funding
  run `npm fund` for details
+ PATH=/usr/src/.npm-global/bin:/usr/bin:/bin:/usr/local/bin
+ cd /usr/src/build
+ rm tools/server/public/index.html.gz
+ cd tools/server/webui
+ aikido-npm ci --ignore-scripts

added 547 packages, and audited 548 packages in 26s

218 packages are looking for funding
  run `npm fund` for details

7 vulnerabilities (1 low, 1 moderate, 5 high)

To address all issues, run:
  npm audit fix

Run `npm audit` for details.
+ aikido-npm audit --audit-level=critical fix

added 12 packages, removed 1 package, changed 8 packages, and audited 559 packages in 27s

227 packages are looking for funding
  run `npm fund` for details

# npm audit report

cookie  <0.7.0
cookie accepts cookie name, path, and domain with out of bounds characters - https://github.com/advisories/GHSA-pxg6-pf52-xh8x
fix available via `npm audit fix --force`
Will install @sveltejs/kit@0.0.30, which is a breaking change
node_modules/cookie
  @sveltejs/kit  >=1.0.0-next.0
  Depends on vulnerable versions of cookie
  node_modules/@sveltejs/kit
    @sveltejs/adapter-static  >=1.0.0-next.0
    Depends on vulnerable versions of @sveltejs/kit
    node_modules/@sveltejs/adapter-static
    runed  >=0.32.0
    Depends on vulnerable versions of @sveltejs/kit
    node_modules/bits-ui/node_modules/runed
      bits-ui  >=2.11.8
      Depends on vulnerable versions of runed
      Depends on vulnerable versions of svelte-toolbelt
      node_modules/bits-ui
      svelte-toolbelt  >=0.10.6
      Depends on vulnerable versions of runed
      node_modules/bits-ui/node_modules/svelte-toolbelt

6 low severity vulnerabilities

To address issues that do not require attention, run:
  npm audit fix

To address all issues (including breaking changes), run:
  npm audit fix --force
ℹ Safe-chain: Some package versions were suppressed due to minimum age requirement.
  To disable this check, use: --safe-chain-skip-minimum-package-age
+ du -sh node_modules
407M	node_modules
+ npm run build

> webui@1.0.0 build
> vite build && ./scripts/post-build.sh

▲ [WARNING] Cannot find base config file "./.svelte-kit/tsconfig.json" [tsconfig.json]

    tsconfig.json:2:12:
      2 │   "extends": "./.svelte-kit/tsconfig.json",
        ╵              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

│
▲  Failed to load preset:
│  "file:/usr/src/build/tools/server/webui/node_modules/@storybook/builder-vite/dist/preset.js"
│
■  Error [ERR_MODULE_NOT_FOUND]: Cannot find module
│  './node_modules/@storybook/builder-vite/dist/preset.js' imported from
│  ./node_modules/storybook/dist/_node-chunks/chunk-OXJURHEN.js
│  at finalizeResolution (node:internal/modules/esm/resolve:274:11)
│  at moduleResolve (node:internal/modules/esm/resolve:859:10)
│  at defaultResolve (node:internal/modules/esm/resolve:983:11)
│  at nextResolve (node:internal/modules/esm/hooks:748:28)
│  at o
│  (file://./node_modules/@tailwindcss/node/dist/esm-cache.loader.mjs:1:69)
│  at nextResolve (node:internal/modules/esm/hooks:748:28)
│  at Hooks.resolve (node:internal/modules/esm/hooks:240:30)
│  at handleMessage (node:internal/modules/esm/worker:199:24)
│  at Immediate.checkForMessages
│  (node:internal/modules/esm/worker:141:28)
│  at process.processImmediate (node:internal/timers:485:21)
vite v7.2.2 building ssr environment for production...
transforming...
DEPRECATION WARNING [import]: Sass @import rules are deprecated and will be removed in Dart Sass 3.0.0.

More info and automated migrator: https://sass-lang.com/d/import

   ╷
17 │ @import 'katex/src/styles/katex.scss';
   │         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
   ╵
    src/styles/katex-custom.scss 17:9  root stylesheet

DEPRECATION WARNING [import]: Sass @import rules are deprecated and will be removed in Dart Sass 3.0.0.

More info and automated migrator: https://sass-lang.com/d/import

  ╷
2 │ @import "./fonts.scss";
  │         ^^^^^^^^^^^^^^
  ╵
    node_modules/katex/src/styles/katex.scss 2:9  @import
    src/styles/katex-custom.scss 17:9             root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.append instead.

More info and automated migrator: https://sass-lang.com/d/import

  ╷
9 │         $src: append($src, url('#{$font-folder}/KaTeX_#{$family}-#{$family-suffix}.woff2') format('woff2'), comma);
  │               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  ╵
    node_modules/katex/src/styles/fonts.scss 9:15   generate-src()
    node_modules/katex/src/styles/fonts.scss 42:11  font-face()
    node_modules/katex/src/styles/fonts.scss 52:1   @import
    node_modules/katex/src/styles/katex.scss 2:9    @import
    src/styles/katex-custom.scss 17:9               root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.length instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
344 │         @for $from from 1 through length($sizes) {
    │                                   ^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 344:35  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.length instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
345 │             @for $to from 1 through length($sizes) {
    │                                     ^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 345:37  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.nth instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
348 │                     font-size: calc((nth($sizes, $to) / nth($sizes, $from)) * 1em);
    │                                      ^^^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 348:38  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.nth instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
348 │                     font-size: calc((nth($sizes, $to) / nth($sizes, $from)) * 1em);
    │                                                         ^^^^^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 348:57  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

node_modules/@sveltejs/kit/src/runtime/client/client.js (333:26): "fork" is not exported by "node_modules/svelte/src/index-server.js", imported by "node_modules/@sveltejs/kit/src/runtime/client/client.js".
"default" is imported from external module "highlight.js" but never used in "src/lib/components/app/misc/SyntaxHighlightedCode.svelte".
✓ 4621 modules transformed.
Export "getJsonHeaders" of module "src/lib/utils/api-headers.ts" was reexported through module "src/lib/utils/index.ts" while both modules are dependencies of each other and will end up in different chunks by current Rollup settings. This scenario is not well supported at the moment as it will produce a circular dependency between chunks and will likely lead to broken execution order.
Either change the import in "src/lib/services/models.ts" to point directly to the exporting module or reconfigure "output.manualChunks" to ensure these modules end up in the same chunk.
Export "getJsonHeaders" of module "src/lib/utils/api-headers.ts" was reexported through module "src/lib/utils/index.ts" while both modules are dependencies of each other and will end up in different chunks by current Rollup settings. This scenario is not well supported at the moment as it will produce a circular dependency between chunks and will likely lead to broken execution order.
Either change the import in "src/lib/services/chat.ts" to point directly to the exporting module or reconfigure "output.manualChunks" to ensure these modules end up in the same chunk.
rendering chunks...
│
▲  Failed to load preset:
│  "file:/usr/src/build/tools/server/webui/node_modules/@storybook/builder-vite/dist/preset.js"
│
■  Error [ERR_MODULE_NOT_FOUND]: Cannot find module
│  './node_modules/@storybook/builder-vite/dist/preset.js' imported from
│  ./node_modules/storybook/dist/_node-chunks/chunk-OXJURHEN.js
│  at finalizeResolution (node:internal/modules/esm/resolve:274:11)
│  at moduleResolve (node:internal/modules/esm/resolve:859:10)
│  at defaultResolve (node:internal/modules/esm/resolve:983:11)
│  at nextResolve (node:internal/modules/esm/hooks:748:28)
│  at o
│  (file://./node_modules/@tailwindcss/node/dist/esm-cache.loader.mjs:1:69)
│  at nextResolve (node:internal/modules/esm/hooks:748:28)
│  at Hooks.resolve (node:internal/modules/esm/hooks:240:30)
│  at handleMessage (node:internal/modules/esm/worker:199:24)
│  at Immediate.checkForMessages
│  (node:internal/modules/esm/worker:141:28)
│  at process.processImmediate (node:internal/timers:485:21)
vite v7.2.2 building client environment for production...
transforming...
DEPRECATION WARNING [import]: Sass @import rules are deprecated and will be removed in Dart Sass 3.0.0.

More info and automated migrator: https://sass-lang.com/d/import

   ╷
17 │ @import 'katex/src/styles/katex.scss';
   │         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
   ╵
    src/styles/katex-custom.scss 17:9  root stylesheet

DEPRECATION WARNING [import]: Sass @import rules are deprecated and will be removed in Dart Sass 3.0.0.

More info and automated migrator: https://sass-lang.com/d/import

  ╷
2 │ @import "./fonts.scss";
  │         ^^^^^^^^^^^^^^
  ╵
    node_modules/katex/src/styles/katex.scss 2:9  @import
    src/styles/katex-custom.scss 17:9             root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.append instead.

More info and automated migrator: https://sass-lang.com/d/import

  ╷
9 │         $src: append($src, url('#{$font-folder}/KaTeX_#{$family}-#{$family-suffix}.woff2') format('woff2'), comma);
  │               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  ╵
    node_modules/katex/src/styles/fonts.scss 9:15   generate-src()
    node_modules/katex/src/styles/fonts.scss 42:11  font-face()
    node_modules/katex/src/styles/fonts.scss 52:1   @import
    node_modules/katex/src/styles/katex.scss 2:9    @import
    src/styles/katex-custom.scss 17:9               root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.length instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
344 │         @for $from from 1 through length($sizes) {
    │                                   ^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 344:35  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.length instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
345 │             @for $to from 1 through length($sizes) {
    │                                     ^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 345:37  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.nth instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
348 │                     font-size: calc((nth($sizes, $to) / nth($sizes, $from)) * 1em);
    │                                      ^^^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 348:38  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

DEPRECATION WARNING [global-builtin]: Global built-in functions are deprecated and will be removed in Dart Sass 3.0.0.
Use list.nth instead.

More info and automated migrator: https://sass-lang.com/d/import

    ╷
348 │                     font-size: calc((nth($sizes, $to) / nth($sizes, $from)) * 1em);
    │                                                         ^^^^^^^^^^^^^^^^^^
    ╵
    node_modules/katex/src/styles/katex.scss 348:57  @import
    src/styles/katex-custom.scss 17:9                root stylesheet

node_modules/@sveltejs/kit/src/runtime/client/client.js (332:26): "fork" is not exported by "node_modules/svelte/src/index-client.js", imported by "node_modules/@sveltejs/kit/src/runtime/client/client.js".
✓ 5356 modules transformed.
rendering chunks...
computing gzip size...
.svelte-kit/output/client/_app/version.json                             0.03 kB │ gzip:     0.05 kB
.svelte-kit/output/client/.vite/manifest.json                           0.33 kB │ gzip:     0.19 kB
.svelte-kit/output/client/_app/immutable/assets/style.sWIT1xAp.css    482.64 kB │ gzip:   286.19 kB
.svelte-kit/output/client/_app/immutable/bundle.DxpKjihs.js         3,777.47 kB │ gzip: 1,118.20 kB

(!) Some chunks are larger than 3072 kB after minification. Consider:
- Using dynamic import() to code-split the application
- Use build.rollupOptions.output.manualChunks to improve chunking: https://rollupjs.org/configuration-options/#output-manualchunks
- Adjust chunk size limit for this warning via build.chunkSizeWarningLimit.
✓ built in 11.56s
.svelte-kit/output/server/.vite/manifest.json                                              5.75 kB
.svelte-kit/output/server/_app/immutable/assets/style.BvrYRhgP.css                       482.35 kB
.svelte-kit/output/server/chunks/environment.js                                            0.07 kB
.svelte-kit/output/server/chunks/api-key-validation.js                                     0.14 kB
.svelte-kit/output/server/entries/pages/_page.ts.js                                        0.17 kB
.svelte-kit/output/server/entries/pages/chat/_id_/_page.ts.js                              0.18 kB
.svelte-kit/output/server/chunks/server.js                                                 0.22 kB
.svelte-kit/output/server/chunks/browser-only.js                                           0.31 kB
.svelte-kit/output/server/internal.js                                                      0.35 kB
.svelte-kit/output/server/chunks/utils.js                                                  0.63 kB
.svelte-kit/output/server/entries/pages/_page.svelte.js                                    0.98 kB
.svelte-kit/output/server/entries/pages/chat/_id_/_page.svelte.js                          1.00 kB
.svelte-kit/output/server/chunks/label.js                                                  2.21 kB
.svelte-kit/output/server/chunks/exports.js                                                3.06 kB
.svelte-kit/output/server/chunks/supported-file-types.js                                   4.32 kB
.svelte-kit/output/server/chunks/internal.js                                               6.02 kB
.svelte-kit/output/server/chunks/shared.js                                                 7.48 kB
.svelte-kit/output/server/remote-entry.js                                                  8.29 kB
.svelte-kit/output/server/entries/pages/_error.svelte.js                                   8.43 kB
.svelte-kit/output/server/chunks/index.js                                                 21.41 kB
.svelte-kit/output/server/entries/pages/_layout.svelte.js                                 34.19 kB
.svelte-kit/output/server/index.js                                                        56.23 kB
.svelte-kit/output/server/chunks/SyntaxHighlightedCode.svelte_svelte_type_style_lang.js   86.69 kB
.svelte-kit/output/server/chunks/DialogConfirmation.js                                   109.90 kB
.svelte-kit/output/server/chunks/ServerLoadingSplash.js                                  217.27 kB
✓ built in 21.04s

Run npm run preview to preview your production build locally.

> Using @sveltejs/adapter-static
Overwriting ../public/index.html with fallback page. Consider using a different name for the fallback.
  Wrote site to "../public"
  ✔ done
✓ Inlined favicon.svg as base64 data URL
✓ Created index.html.gz
2026-01-24 04:50:31 +03:00
Vitaly Chikunov
42d6c582aa .gear/generate: Add proxy and checklist support
Safe-chain: Unhandled promise rejection: Error: Error parsing malware database: Invalid response body while trying to fetch https://malware-list.aikido.dev/malware_predictions.json: read ETIMEDOUT

Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
2026-01-24 04:50:31 +03:00
Vitaly Chikunov
bfba6a3251 Merge signed commit 'b7819' into sisyphus
Merge-Strategy: theirs
Diff-After-Merge: 1 file changed, 6 insertions(+)

# gpg: Signature made Fri Jan 23 19:22:34 2026 MSK
# gpg:                using RSA key B5690EEEBB952194
# gpg: Good signature from "GitHub <noreply@github.com>" [unknown]
2026-01-24 01:48:29 +00:00
Georgi Gerganov
557515be1e
graph : utilize ggml_build_forward_select() to avoid reallocations (#18898)
* graph : avoid branches between embedding and token inputs

* models : make deepstack graphs (e.g. Qwen3 VL) have constant topology

* ci : enable -DGGML_SCHED_NO_REALLOC=ON for server CI

* cont : pad token embeddings to n_embd_inp
2026-01-23 18:22:34 +02:00
Neo Zhang
cb6caca191
[SYCL] use malloc to support both iGPU and dGPU in same time (#18992)
* use malloc to support both iGPU and dGPU in same time

* support windows

---------

Co-authored-by: Neo Zhang Jianyu <jianyu.zhang@intel.com>
2026-01-23 20:54:10 +08:00
Xuan-Son Nguyen
b5b8fa1c8b
chat : fix translategemma crash on common_chat_format_example (#19019) 2026-01-23 12:03:42 +01:00
Daniel Bevenius
a14b960bc7
model-conversion : use BUILD_DIR variable in all scripts (#19015)
This commit modifies all the utility scripts to use an optional
BUILD_DIR variable/argument to specify the build directory.

The motivation for this is that Commit
3d55846a5c ("model-conversion : add
BUILD_DIR variable to run-converted-model scripts") introduced this
variable to the causal and embeddings scripts, but I missed the scripts
in the utils directory.
2026-01-23 09:01:36 +01:00
Alberto Cabrera Pérez
091a46cb8d
ggml-cpu: aarm64: q5_K repack gemm and gemv (and generic) implementations (i8mm) (#18860)
* Boilerplate for q5_Kx8 REPACK on ARM and fallback

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>

* Implements make_block_q5_Kx8 by extending make_block_q4_Kx8

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>

* q5_K repack gemm and gemv generics

* Gemm and Gemv ARM implementations (i8mm)

* Improved qh manipulation looking at non-repack vec_dot implementation

* Full unroll

* Apply Q5_K Gemv vand and vshl optimizations to gemm. Improve comments.

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>

* Fix wrong fallback definitions of Q5_K

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>

* Fixed comments. Reverted unnecessary formatting

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>

* Fixed typo in generic definitions

* Switching AND + Shift with Shift Insert. Better op interleaving.

* Vectorize + unroll the block scales

* Apply gemm optimizations to gemv

* Improve bias calculation

---------

Signed-off-by: Alberto Cabrera <alberto.cabrera@liquid.ai>
2026-01-23 09:55:08 +02:00
Aldehir Rojas
a3e812811d
cli : load parser definition (#19031)
* cli : load parser definition

* cont : only unload if a parser is defined
2026-01-22 20:31:22 -06:00
Xuan-Son Nguyen
51fa458a92
server : support preserving reasoning_content in assistant message (#18994)
* support reasoning_content input

* report template caps to webui

* add docs

* rm commented code
2026-01-22 21:30:06 +01:00
Georgi Gerganov
a5eaa1d6a3
mla : make the V tensor a view of K (#18986)
* mla : pass V as a view of K to the FA op

* cuda : adjust mla logic to new layout

* kv-cache : fix rope shift

* tests : remove comment

* cuda : fix reusable_cutoff

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

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2026-01-22 22:09:01 +02:00
Johannes Gäßler
e2baf02162
CUDA: fix alignment check for FA (#19023) 2026-01-22 20:39:25 +01:00
Aman Gupta
e34d6d03b2
convert_hf_to_gguf.py: refactor modify_tensors to call super (#18866) 2026-01-23 02:58:07 +08:00
lhez
9c96465f99
opencl: enable the general fp mm for non-cont input and as a fallback for specialized kqv kernel for adreno (#18970)
* opencl: add `copy_to_contiguous` and utilize mm kernels

* opencl: only copy to cont for f32 and f16 tensors

* opencl: use cont mm for fallback when dst is large

* opencl: use nb local to copy-to-cont

* opencl: use local offset as well
2026-01-22 10:29:25 -08:00
Xuan-Son Nguyen
4e595b250a
server: do not log certain endpoints (avoid log spam) (#19028) 2026-01-22 19:24:37 +01:00
Georgi Gerganov
0e4ebeb057
quant : manual overrides of tensor types take precedence (#18952) 2026-01-22 16:17:06 +02:00
Aaron Teo
8b30840703
release: update github api (#19022) 2026-01-22 21:38:02 +08:00
Xuan-Son Nguyen
9eb5bfec1a
mtmd : update docs to use llama_model_n_embd_inp (#18999) 2026-01-22 14:36:32 +01:00
손희준
c6926d1d95
server: Reorder methods in server-task.cpp (#19016)
* Move `task_result_state::update_chat_msg` to match with header

* Move `server_task_result_cmpl_partial::to_json_anthropic()` to match with header

---------

Co-authored-by: openingnow <>
2026-01-22 14:36:04 +01:00
Aman Gupta
b70d251076
CUDA: add gqa_ratio 4 for GLM 4.7 flash (#18953) 2026-01-22 18:51:53 +08:00
shaofeiqi
5516b9c16a
opencl: add TRI op support (#18979) 2026-01-21 22:05:54 -08:00
Aleksei Nikiforov
94242a62c0
ggml-zdnn : mark zDNN buffers as non-host (#18967)
While buffers reside in host memory,
additional transformation is needed to use buffers with zDNN.

Fixes #18848
2026-01-22 01:16:21 +01:00
Pádraic Slattery
6b99a223e3
ci : update GitHub Actions versions [no ci] (#18935) 2026-01-22 00:57:18 +01:00
Mariusz Woloszyn
77078e80e5
convert : add Devstral-2 (Ministral3ForCausalLM) arch (#18972)
* Add Ministral3ForCausalLM architeture

This adds support for newer architectres like Devstral-2

* removed blank line found after function decorator

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

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-01-22 00:55:55 +01:00
Piotr Wilkin (ilintar)
c301172f66
jinja: support none|string (#18995)
* jinja: support none|string

* Update common/jinja/value.cpp

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

* Update tests/test-jinja.cpp

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

* Add as_string()

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2026-01-21 19:24:37 +01:00
Hendrik Erz
3802d3c78f
fix: Use tabular-nums for chat message statistics (#18915)
* fix: Use `tabular-nums` for chat message statistics

* fix: Rebuild WebUI
2026-01-21 18:46:01 +01:00
Daniel Bevenius
9da3dcd753
llama : clarify nemotron-h.cpp comment about RoPE [no ci] (#18997)
This commit removes the mention of RoPE in the comment for the Q and K
computation as RoPE is not applied.
2026-01-21 18:31:34 +01:00
Jeff Bolz
bd544c94a3
vulkan: Remove transfer_ctx, do everything in compute_ctx. (#18945)
* vulkan: Remove transfer_ctx, do everything in compute_ctx.

We had a bug where a set_tensor_async (using transfer_ctx) didn't get
submitted before the graph_compute (using compute_ctx) that came after
it. To avoid this sort of issue, just do everything in compute_ctx.

Remove transfer_cmd_pool, which was already unused.

* fix crash with perf logger
2026-01-21 18:01:40 +01:00
Adrien Gallouët
14be5a39b1
common : improve error message when HTTPS is missing but required (#18987)
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2026-01-21 17:58:38 +01:00
손희준
fbbf3ad190
server: /v1/responses (partial) (#18486)
* from previous PR

* Make instruction(system) as first message

* Convert [input_message] (text/image/file)

* Rename convert_responses_to_chatcmpl(body) -> response_body

* Initial tool call support

* Erase instructions field from chatcmpl body

* Feed reasoning texts to chat template

* Use std::vector instead of opaque json array

* Make output_item.added events consistent

* Move `server_task_result_cmpl_partial::update` from header to source

* Match ID of output_item.added and .done events

* Add function_call only if there is no "fc_" prefix

* Add function call output at non-streaming API

* Test if ID is persistent

* Add doc

* Fix style - use trailing comma

* Rewrite state management

* catch up with upstream/master

* Fix style - "type" is the first item of SSE data

* Explicitly check "instructions" from response_body

* Make lambdas static

* Check if reasoning content exists

* Add `oai_resp_id` to task_result_state(also initialized at ctor), server_task_result_cmpl_partial, and server_task_result_cmpl_final

* Reject `input_file` since it is not supported by chatcmpl

* Add "fc_" prefix to non-straming function call id as coderabbit pointed out

---------

Co-authored-by: openingnow <>
2026-01-21 17:47:23 +01:00
Jeff Bolz
33f890e579
vulkan: support flash attention GQA/split_k with small batches (#18938) 2026-01-21 17:43:43 +01:00
Masato Nakasaka
067b8d7af3
Revert "vulkan: force full subgroups for flash attention to fix intel subgroup crash (#17356)" (#18831)
This reverts commit 980b7cd17e.
2026-01-21 17:13:43 +01:00
Jeff Bolz
50b7f076a5
vulkan: Use mul_mat_vec_id for small values of n (#18918)
Change ggml_vk_mul_mat_vec_id_q_f16 to loop over the batch dimension and
update the indexing calculations in get_offsets.

Mat-vec is faster than mat-mat for small values of n. We don't get the same
reuse of the weights as in the non-ID path, but with this the cost is linear
in n rather than n>1 being far slower than n==1.
2026-01-21 16:22:02 +01:00
Tarek Dakhran
ad8d85bd94
memory : add llama_memory_hybrid_iswa (#18601)
* memory : add llama_memory_hybrid_iswa

* Update src/llama-memory-hybrid-iswa.cpp

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

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-01-21 14:30:23 +02:00
Piotr Wilkin (ilintar)
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
Matthieu Coudron
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
Oliver Simons
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
Adrien Gallouët
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
Xuan-Son Nguyen
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
Oliver Simons
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
Adrien Gallouët
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
Georgi Gerganov
271191906c
metal : enable FA for MLA heads (#18950) 2026-01-20 12:21:28 +02:00
Daniel Bevenius
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
Xuan-Son Nguyen
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
ddh0
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
Sigbjørn Skjæret
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
Sigbjørn Skjæret
4037093c66
ci : run test-jinja -py on high perf [no ci] (#18916) 2026-01-19 20:29:15 +01:00
Lennart Austenfeld
18361c579c
server: fix memory reservations in populate_token_probs (#18787) 2026-01-19 19:13:31 +01:00
Georgi Gerganov
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
Daniel Bevenius
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