Skip to content

Oom/fixing#4

Draft
peterchen-intel wants to merge 90 commits into
masterfrom
oom/fixing
Draft

Oom/fixing#4
peterchen-intel wants to merge 90 commits into
masterfrom
oom/fixing

Conversation

@peterchen-intel
Copy link
Copy Markdown
Owner

Details:

  • item1
  • ...

Tickets:

  • ticket-id

AI Assistance:

  • AI assistance used: yes
  • If yes, summarize how AI was used and what human validation was performed (build/tests/manual checks).

causing it to be skipped on MTL-class iGPU (12.70.x, XeHPG, no DPAS).
This left raw FP32 weight-decompression chains that overwhelmed
propagate_constants with ~56 GB of constant-folding memory.

Root cause of inference failure: moe_3gemm_swiglu_opt uses oneDNN
internally (onednn_linear for gate/up/down matrix multiplications).
OneDNN requires an in-order OCL queue.  MTL uses out-of-order queue
by default because use_onednn is false when supports_immad=false.

Fix:
  three MoE transformation passes (FuseVectorizedMOE3GEMM,
  ConvertMOEToMOECompressed, FuseMOE3GemmCompressed) run on all
  architectures.  FuseMOE3GemmCompressed creates MOE3GemmFusedCompressed
  which the OCL moe_3gemm_swiglu_opt kernel executes.
- Detect MOE3GemmFusedCompressed in apply_model_specific_options and
  force use_onednn=true so finalize_impl sets queue_type=in_order,
  satisfying the oneDNN in-order queue requirement.
- Fix moe_gather validate_impl to accept rank-2 input for models where
  the batch dimension is pre-flattened (Qwen3-style).
- Re-apply iGPU transfer skip (usm_shared -> usm_device) in
  network.cpp and program.cpp for integrated GPUs where both allocation
  types share system DRAM (xe2+ or 12.7x-class MTL/ARL-S).

Tested on machine (GPU uArch 12.70.4 / XeHPG / System memory 64 GB): model loads in 14 s,
generates meaningful tokens, Unevictable stays below 120 MB.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Signed-off-by: Chen Peter <peter.chen@intel.com>
peterchen-intel and others added 6 commits April 14, 2026 05:21
Instead of setting use_onednn=true when MOE3GemmFusedCompressed is detected,
set m_queue_type=in_order directly.  This is more precise: the only requirement
is an in-order OCL command queue (for onednn_linear in moe_3gemm_swiglu_opt.cpp),
not full oneDNN enablement for the whole model.

Leaving use_onednn=false on non-systolic hardware (MTL, 12.70.x) ensures that
oneDNN implementations for FC, convolution, GEMM etc. are not activated on
hardware without DPAS units.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The MTL-class (12.70.x) iGPU has a separate GPU L3 cache from the CPU,
so copying usm_shared -> usm_device does improve GPU access performance.
Reverts the MTL condition added in the prior fix commit, keeping only
the original xe2+ integrated GPU skip (which has true unified memory).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
In the correct fix, GEMM3_SWIGLU (Qwen3) always goes through
FuseMOE3GemmCompressed -> MOE3GemmFusedCompressed, which creates a
single fused primitive with no standalone moe_gather node.

The rank-2 accept was only needed during an intermediate broken debug
state where FuseMOE3GemmCompressed was wrongly blocked.  moe_gather is
only used by GEMM2_BIAS_SWIGLU_CLAMP models, whose input is rank-3.

Restore original: input_pshapes.rank() != 3 || input_pshapes[2].is_dynamic()

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The previous fix set m_queue_type = in_order directly in
apply_model_specific_options, but this left m_use_onednn = false.
On non-systolic hardware (supports_immad=false), program.cpp only
calls lo.enable_onednn_for<lstm_seq/gru_seq>() (making the
onednn_impls_optimization_attribute non-empty, which triggers
create_onednn_engine() in select_preferred_formats.cpp) when
use_onednn=true.  With use_onednn=false, the engine is never
initialized, causing moe_3gemm_fused_compressed to crash at
inference time with 'oneDNN engine not initialized'.

Fix: set m_use_onednn = true (not queue_type) when
MOE3GemmFusedCompressed is detected. finalize_impl then sets
queue_type = in_order because use_onednn=true, and the
create_onednn_engine() call is correctly triggered.

This is safe on non-systolic hardware: FuseVectorizedFC (systolic
FC) is gated independently on supports_immad, so no systolic ops
are introduced by enabling use_onednn for the MoE path.

Verified: all 3 prompts pass with correct output on MTL iGPU
(GPU_UARCH_VERSION=12.70.4, supports_immad=false).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
paged_attention_opt__multi_tokens allocates a tmp_out scratch buffer sized
total_tokens * heads_num * v_head_size * num_of_partitions * sizeof(float).
For Qwen3-30B with chunk_size=4096 and 8K KV context this is 2 GB per layer.
With 48 layers all executing sequentially, this totalled 96 GB of demand-paged
USM device allocation. On Intel iGPU (ARLS, i915 driver), the driver pins the
entire allocation as Unevictable on first GPU access regardless of pages touched,
causing CL_OUT_OF_RESOURCES on a 31 GB machine.

Root cause: can_share_internal_buffer(false) in paged_attention_node unconditionally
blocked the memory pool for ALL internal buffers. This was added in PR openvinotoolkit#33204 to
prevent CPU/GPU races on lockable buffers (blocks_indexes_start/end,
blocked_gws_subseq_mapping) written by prepare_internal_buffers(). However it also
blocked pool reuse for non-lockable GPU-only buffers (exp_sums, max_logits, tmp_out)
which are safe to share across sequential layers.

Fix:
- Remove can_share_internal_buffer(false) from paged_attention_node; per-buffer
  lockability already tracked via BufferDescriptor::m_lockable.
  so CPU-written (lockable=true, usm_host) buffers remain non-shareable while
  GPU-only (lockable=false, usm_device) buffers can be reused from the pool.
- In allocate_internal_buffers(): pass buffer_descs[i].m_lockable to the call
  (previously dropped, causing wrong alloc type on initial allocation).

Result: 48 layers share one 2 GB tmp_out buffer instead of allocating 48 separate
2 GB buffers. Peak Unevictable drops from OOM crash (~28+ GB) to ~18.9 GB on ARLS
(Intel Arc 8086:7d67, Arrow Lake-S iGPU, 31 GB).

Verified: Qwen3-30B-A3B-Instruct-2507-int4-ov with chunk_size=4096, 8K prompt,
ContinuousBatching on ARLS completes successfully with exit code 0 and 20 coherent
output tokens. Not affected on ARLH (supports_immad=true takes micro_sdpa path
which does not allocate tmp_out at all).

Signed-off-by: Chen Peter <peter.chen@intel.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
peterchen-intel and others added 21 commits April 21, 2026 03:29
## Summary

Make file-based model cache write failures recoverable and keep cache
entries consistent.

Before this change, a cache write failure during model export could
leave an invalid cache entry and break subsequent cache use.

After this change:
- file-based cache writes either complete successfully or the failed
cache entry is removed
- recoverable cache export failures no longer fail model compilation
- the cache write path stays simple and writes directly to the final
blob path

### Details:
- File-based cache writes directly to the final `.blob` path.
- Repeated writes to the same cache entry overwrite the existing blob.
- Existing blob permissions are temporarily relaxed before overwrite,
then restored to read-only after a successful store.
- Cache export failures are handled in `CoreImpl`, which removes the
failed cache entry on caching errors.
- Stream write failures remain recoverable and do not fail model
compilation.
- GPU export failures that surface as `ov::Exception` are also treated
as recoverable cache export failures.
- Other unexpected exceptions are still propagated.

### Result:
- before: a cache write failure could leave a broken cache entry and
break subsequent cache use
- after: failed cache writes are discarded, cache state remains usable,
and subsequent runs continue normally

### Tickets:
- CVS-104958

### AI Assistance:
- *AI assistance used: yes*
### Details:
 - fixing segmentation fault for convolution tests in Level Zero runtime
- changing ```ASSERT_EQ(1,1); return;``` to ```GTEST_SKIP()``` for
skipping in unit tests
 - adding ```get_context()``` method to ```ze_stream```
- ```GPU_DEBUG_INFO``` for ```ze_stream::flush()``` method in Level Zero
- avoiding program termination while stack unwinding (in
```OV_ZE_EXPECT```)

### Tickets:
 - CVS-179218

### AI Assistance:
 - *AI assistance used: yes*
openvinotoolkit#35419)

### Details:
- Fix implicit narrowing conversions (C4244) in GPU plugin to prepare
for /wd4244 removal (BinSkim BA2007 compliance)
- loop.hpp: Change constructor param from int64_t to int32_t with
OPENVINO_ASSERT range check at caller
- memory_pool.cpp/hpp: Change debug accumulators from float to size_t
(float loses precision above 16MB)
- swiglu_with_clamp.hpp / swiglu_kernel_base.h: Correct setter and
fuse_params types to float to match member types
- paged_attention_gen.cpp: Use float sqrt path instead of double for
scale_factor
- custom_gpu_primitive.hpp: Add static_cast<int> in GetDim macro to
prevent int64_t→int narrowing

### Tickets:
 - CVS-185007

### AI Assistance:
 - AI assistance used: yes
- GitHub Copilot used for code analysis (identifying root causes of each
warning, evaluating fix options for type safety vs. serialization
compatibility), generating fix code, and verifying build results. All
changes were human-reviewed for correctness, validated via local /WX
incremental build, and confirmed to have no performance or accuracy
impact.
…path (openvinotoolkit#35408)

### Details:
- _tensorflow_lite/src/frontend.cpp_ methods pass
`std::filesystem::path` directly instead of extracting `.native()`
- Replaces `std::string/std::wstring` path parameters with
`std::filesystem::path` in `GraphIteratorFlatBuffer` constructor and
`is_supported()` method

### Tickets:
- part of
[CVS-146661](https://jira.devtools.intel.com/browse/CVS-146661)

### AI Assistance:
 - yes, codebase analysis
…t#35324)

### Details:
 - Use minimal reduction workspace for TopK=1 (avoid full sort buffers)
 - Add unit tests with large random inputs and dynamic shapes
 - No behavior change for TopK > 1
 - Improves memory usage and reduces overhead for argmax/argmin.

### Tickets:
 - *CVS-184749*

### AI Assistance:
 - *AI assistance used: no *

Co-authored-by: Chon Ming Lee <chon.ming.lee@intel.com>
…#34566)

### Details:
- Removed broken maintainer links and outdated file links in
`src/plugins/intel_gpu/README.md` and `src/plugins/intel_cpu/README.md`.
- I noticed the `.take` bot didn't assign the issue to me, possibly due
to the label. Since this is a quick and straightforward documentation
fix, I went ahead and prepared the PR to save the maintainers' time.

### Tickets:
- Resolves openvinotoolkit#34556

### AI Assistance:
- AI assistance used: no

---------

Co-authored-by: Pavel Durandin <pavel.durandin@intel.com>
### Details:
- *Fix dynamic bounds by forcing to create a new level zero tensor with
max size and do a memcpy all the time*
- *Throw if the dimension of the dynamic rank is different than the
expected one for the indexes that aren't dynamic.*

### Tickets:
 - *N/A*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: Bogdan Pereanu <bogdan.pereanu@intel.com>
…t#35356)

### Details:
- Removed from _frontend.cpp_ (`supported_impl`, `load_impl`) most
usages of `path::native()` and `ov::util::path_to_string()`
- Replaced template-based and `std::string` path parameters with
`std::filesystem::path` in `CheckpointV1Reader`, `GraphIteratorProto`,
`GraphIteratorProtoTxt`, and `GraphIteratorMeta`.
- `ov::util::path_to_string(`) is used for `GraphIteratorSavedModel`
second argument "tags", which originally can be passed into `.load()`
method as `std::string` or `std::wstring`
- `m_shard_names` is used for error messages - not a path for I/O
operations so I kept it as `std::string`.

These changes make the codebase more robust, easier to maintain, and
better aligned with modern C++ practices.

### Tickets:
 - [CVS-184167](https://jira.devtools.intel.com/browse/CVS-184167)

### AI Assistance:
 - AI assistance used: yes
 - Codebase analysis
### Details:
 - *coverity scan issue of underflow and overflow fix*

### Tickets:
 - *CVS-185067*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*
…rmalizationFusion (openvinotoolkit#35195)

### Details:

Some frameworks decompose GroupNormalization via InstanceNormalization
and reshape the grouped tensor directly to 4D {N,G,-1,1} with a trailing
unit dimension. MVN then reduces over axes {2,3} instead of axis {2} as
in the existing 3D pattern {N,G,-1}. This is mathematically equivalent
since the trailing dimension is always 1.

Changes:
- Extend `pre_mvn_reshape_m` predicate from `rank_equals(3)` to
`is_rank_3_or_4` so the pattern matcher accepts both 3D and 4D reshapes
before MVN.
- Validate the trailing dimension is 1 for the 4D variant.
- Handle MVN axes {2,3} validation for the 4D variant, including
negative axis normalization.
- Allow concrete values (not just -1) in the pre-MVN reshape's third
dimension, since some frameworks emit explicit flattened sizes after
constant folding.
- Add positive tests for both special-marker ({0,G,-1,1}) and
concrete-value 4D shapes, covering f32 and f16.
- Add negative edge-case tests: wrong MVN axes, single axis with 4D
reshape, trailing dim != 1, axes over batch/groups, and three MVN axes.

### Tickets:
- CVS-184129

### AI Assistance:
 - *AI assistance used: yes*
- *VS Code Copilot(Claude Opus 4.6) was used to create code changes, to
execute unit tests.*

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Pavel Durandin <pavel.durandin@intel.com>
…envinotoolkit#35406)

Fix GPU kernel compilation: correct return type in binary search macros.

DECLARE_LOWER/UPPER_BOUND macros declared `Type*` return type but
returned `uint` indices. This caused kernel compilation failure on
certain GPU targets when bucketize uses float input + float16 buckets.

Changed: `inline Type*`-> `inline uint`

Compiler behavior may vary across hardware architectures, which is why
this fix ensures consistent compilation across all targets.

### Tickets:
 - *185056*

### AI Assistance:
 - *AI assistance used: yes
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

Signed-off-by: hyunback <hyunback.kim@intel.com>
### Details:
- *This PR adds overflow-safe validation for large shape/memory
calculations in Tile and CPU blocked memory descriptor paths, to fail
early with clear errors instead of reaching undefined arithmetic
behavior or late allocation failures.*

### Tickets:
 - *CVS-184032*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*
)

### Details:
 - Integrate grouped gemm to support moe_3gemm
- Grouped gemm will be enabled by default, and can be disabled by
MOE_USE_GROUPED_GEMM_PREFILL=0

- Test result if applied DNNL_ARG_HINT_MAX_GROUP_SIZE

<img width="772" height="391" alt="image"
src="https://github.com/user-attachments/assets/71e51cc6-6561-411c-826f-6e0ca8fd9b66"
/>


### Tickets:
 - *CVS-182465*
 - *CVS-179099*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Co-authored-by: Chen Peter <peter.chen@intel.com>
…openvinotoolkit#35423)

### Details:
- _paddle/src/frontend.cpp_ passes std::filesystem::path directly
instead of extracting .native()
- Replaces` std::string/std::wstring` path parameters with
`std::filesystem::path` in `InputModel` constructor, `load_consts()`,
functions (`get_const_path`, `is_pdmodel`, `get_model_path`)
- Removes no longer needed `get_path_sep()` template functions from
paddle_utils.hpp

### Tickets:
- part of
[CVS-146661](https://jira.devtools.intel.com/browse/CVS-146661)

### AI Assistance:
 - yes, codebase analysis
…RoPEFusion (openvinotoolkit#34527)

### Details:
RoteryEmbedding is decomposed in OpenVINO at [1] or [2] using gpt-oss
style, but can't be fused by RoPEFusionGPTOSS[3]. The two main points
are:
1. Split is used in decompose, but VariadicSplit is required in
RoPEFusionGPTOSS.
2. Sub is used in decompose, but add with negative value is required in
RoPEFusionGPTOSS.

Using decomposed graph to excute will impact performance. So we should
make the decomposed graph be fused again.

This PR also extend RoPEFusionGPTOSS to support the last dimension
represented by positive numbers for Concat.

[1]
[https://github.com/openvinotoolkit/openvino/blob/master/src/frontends/onnx/frontend/src/op/com.microsoft/rotary_embedding.cpp#L58](https://github.com/openvinotoolkit/openvino/blob/master/src/frontends/onnx/frontend/src/op/com.microsoft/rotary_embedding.cpp)


[2][https://github.com/openvinotoolkit/openvino/blob/master/src/common/transformations/src/transformations/op_conversions/group_query_attention_decomposition.cpp#L256](https://github.com/openvinotoolkit/openvino/blob/master/src/common/transformations/src/transformations/op_conversions/group_query_attention_decomposition.cpp)


[3][https://github.com/openvinotoolkit/openvino/blob/master/src/common/transformations/src/transformations/common_optimizations/fuse_rotary_positional_embeddings.cpp#L1138](https://github.com/openvinotoolkit/openvino/blob/master/src/common/transformations/src/transformations/common_optimizations/fuse_rotary_positional_embeddings.cpp)

### Jira Ticket:

- https://jira.devtools.intel.com/browse/CVS-182397
…inotoolkit#35421)

Bumps the npm-packages group in /samples/js/node with 2 updates:
[@napi-rs/canvas](https://github.com/Brooooooklyn/canvas) and
[eslint](https://github.com/eslint/eslint).

Updates `@napi-rs/canvas` from 0.1.98 to 0.1.99
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/Brooooooklyn/canvas/releases"><code>@​napi-rs/canvas</code>'s
releases</a>.</em></p>
<blockquote>
<h2>v0.1.99</h2>
<h2>What's Changed</h2>
<ul>
<li>chore(deps): update yarn to v4.14.0 by <a
href="https://github.com/renovate"><code>@​renovate</code></a>[bot] in
<a
href="https://redirect.github.com/Brooooooklyn/canvas/pull/1249">Brooooooklyn/canvas#1249</a></li>
<li>chore(deps): update yarn to v4.14.1 by <a
href="https://github.com/renovate"><code>@​renovate</code></a>[bot] in
<a
href="https://redirect.github.com/Brooooooklyn/canvas/pull/1251">Brooooooklyn/canvas#1251</a></li>
<li>fix: drawImage gray halo on transparent PNG edges with
imageSmoothingEnabled by <a
href="https://github.com/Brooooooklyn"><code>@​Brooooooklyn</code></a>
in <a
href="https://redirect.github.com/Brooooooklyn/canvas/pull/1252">Brooooooklyn/canvas#1252</a></li>
<li>Allow canvas package postinstall script for benchmark CI by <a
href="https://github.com/Claude"><code>@​Claude</code></a> in <a
href="https://redirect.github.com/Brooooooklyn/canvas/pull/1253">Brooooooklyn/canvas#1253</a></li>
</ul>
<h2>New Contributors</h2>
<ul>
<li><a href="https://github.com/Claude"><code>@​Claude</code></a> made
their first contribution in <a
href="https://redirect.github.com/Brooooooklyn/canvas/pull/1253">Brooooooklyn/canvas#1253</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/Brooooooklyn/canvas/compare/v0.1.98...v0.1.99">https://github.com/Brooooooklyn/canvas/compare/v0.1.98...v0.1.99</a></p>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/Brooooooklyn/canvas/blob/main/CHANGELOG.md"><code>@​napi-rs/canvas</code>'s
changelog</a>.</em></p>
<blockquote>
<h2><a
href="https://github.com/Brooooooklyn/canvas/compare/v0.1.98...v0.1.99">0.1.99</a>
(2026-04-18)</h2>
<h3>Bug Fixes</h3>
<ul>
<li>drawImage gray halo on transparent PNG edges with
imageSmoothingEnabled (<a
href="https://redirect.github.com/Brooooooklyn/canvas/issues/1252">#1252</a>)
(<a
href="https://github.com/Brooooooklyn/canvas/commit/a748f3f565cfdfb975e28b5ee13f702000a9059a">a748f3f</a>)</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/Brooooooklyn/canvas/commit/0372a4af68c6fe4f0e872a0dd7874b7c436bc562"><code>0372a4a</code></a>
0.1.99</li>
<li><a
href="https://github.com/Brooooooklyn/canvas/commit/adc6e1719f67cb9c80e8ed3373c06ed9abf9e2e4"><code>adc6e17</code></a>
ci: allow canvas package postinstall script for benchmark CI (<a
href="https://redirect.github.com/Brooooooklyn/canvas/issues/1253">#1253</a>)</li>
<li><a
href="https://github.com/Brooooooklyn/canvas/commit/a748f3f565cfdfb975e28b5ee13f702000a9059a"><code>a748f3f</code></a>
fix: drawImage gray halo on transparent PNG edges with
imageSmoothingEnabled ...</li>
<li><a
href="https://github.com/Brooooooklyn/canvas/commit/d2d01c338f4637c7331cef34878bceb8b9c06a39"><code>d2d01c3</code></a>
chore(deps): update yarn to v4.14.1 (<a
href="https://redirect.github.com/Brooooooklyn/canvas/issues/1251">#1251</a>)</li>
<li><a
href="https://github.com/Brooooooklyn/canvas/commit/1f595c6998a58d04b265c24ba5ce204acfd8d730"><code>1f595c6</code></a>
chore(deps): update yarn to v4.14.0 (<a
href="https://redirect.github.com/Brooooooklyn/canvas/issues/1249">#1249</a>)</li>
<li>See full diff in <a
href="https://github.com/Brooooooklyn/canvas/compare/v0.1.98...v0.1.99">compare
view</a></li>
</ul>
</details>
<br />

Updates `eslint` from 10.2.0 to 10.2.1
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/eslint/eslint/releases">eslint's
releases</a>.</em></p>
<blockquote>
<h2>v10.2.1</h2>
<h2>Bug Fixes</h2>
<ul>
<li><a
href="https://github.com/eslint/eslint/commit/14be92b6d1fa0923b8923830f2208e5e2705b002"><code>14be92b</code></a>
fix: model generator yield resumption paths in code path analysis (<a
href="https://redirect.github.com/eslint/eslint/issues/20665">#20665</a>)
(sethamus)</li>
<li><a
href="https://github.com/eslint/eslint/commit/84a19d2c32255db6b9cfc08644a607aae6d5cb62"><code>84a19d2</code></a>
fix: no-async-promise-executor false positives for shadowed Promise (<a
href="https://redirect.github.com/eslint/eslint/issues/20740">#20740</a>)
(xbinaryx)</li>
<li><a
href="https://github.com/eslint/eslint/commit/af764af0ec38225755fbf8a6f207f0c77b595a8d"><code>af764af</code></a>
fix: clarify language and processor validation errors (<a
href="https://redirect.github.com/eslint/eslint/issues/20729">#20729</a>)
(Pixel998)</li>
<li><a
href="https://github.com/eslint/eslint/commit/e251b89a38280973e468a4a9386c138f4f55d10d"><code>e251b89</code></a>
fix: update eslint (<a
href="https://redirect.github.com/eslint/eslint/issues/20715">#20715</a>)
(renovate[bot])</li>
</ul>
<h2>Documentation</h2>
<ul>
<li><a
href="https://github.com/eslint/eslint/commit/ca92ca0fb4599e8de1e2fb914e695fe7397cbe63"><code>ca92ca0</code></a>
docs: reuse markdown-it instance for markdown filter (<a
href="https://redirect.github.com/eslint/eslint/issues/20768">#20768</a>)
(Amaresh S M)</li>
<li><a
href="https://github.com/eslint/eslint/commit/57d2ee213305cee0cb55ef08e0480b57396269a9"><code>57d2ee2</code></a>
docs: Enable Eleventy incremental mode for watch (<a
href="https://redirect.github.com/eslint/eslint/issues/20767">#20767</a>)
(Amaresh S M)</li>
<li><a
href="https://github.com/eslint/eslint/commit/c1621b915742276e5f4b25efe790ca62296330dc"><code>c1621b9</code></a>
docs: fix typos in code-path-analyzer.js (<a
href="https://redirect.github.com/eslint/eslint/issues/20700">#20700</a>)
(Ayush Shukla)</li>
<li><a
href="https://github.com/eslint/eslint/commit/1418d522d10bde1960f4942afb548bc7160ec49e"><code>1418d52</code></a>
docs: Update README (GitHub Actions Bot)</li>
<li><a
href="https://github.com/eslint/eslint/commit/39771e6e600f0b0617fdeafff6dd07e4211ffde6"><code>39771e6</code></a>
docs: Update README (GitHub Actions Bot)</li>
<li><a
href="https://github.com/eslint/eslint/commit/71e04693def2df57268f08f3072a2749df6bf438"><code>71e0469</code></a>
docs: fix incomplete JSDoc param description in no-shadow rule (<a
href="https://redirect.github.com/eslint/eslint/issues/20728">#20728</a>)
(kuldeep kumar)</li>
<li><a
href="https://github.com/eslint/eslint/commit/22119ceb93e28f62262fc1d98ff1b1442d6e2dbf"><code>22119ce</code></a>
docs: clarify scope of for-direction rule with dead code examples (<a
href="https://redirect.github.com/eslint/eslint/issues/20723">#20723</a>)
(Amaresh S M)</li>
<li><a
href="https://github.com/eslint/eslint/commit/8f3fb77f122a5641d1833cad5d93f3f54fa3be0b"><code>8f3fb77</code></a>
docs: document <code>meta.docs.dialects</code> (<a
href="https://redirect.github.com/eslint/eslint/issues/20718">#20718</a>)
(Pixel998)</li>
</ul>
<h2>Chores</h2>
<ul>
<li><a
href="https://github.com/eslint/eslint/commit/7ddfea9c4f62add1588c5c0b0da568c299246383"><code>7ddfea9</code></a>
chore: update dependency prettier to v3.8.2 (<a
href="https://redirect.github.com/eslint/eslint/issues/20770">#20770</a>)
(renovate[bot])</li>
<li><a
href="https://github.com/eslint/eslint/commit/fac40e1de2ba7646cc7cd2d3f93fbdd1f8819001"><code>fac40e1</code></a>
ci: bump pnpm/action-setup from 5.0.0 to 6.0.0 (<a
href="https://redirect.github.com/eslint/eslint/issues/20763">#20763</a>)
(dependabot[bot])</li>
<li><a
href="https://github.com/eslint/eslint/commit/7246f923332522d8b3d46b6ee646fce88535f3fb"><code>7246f92</code></a>
test: add tests for SuppressionsService.load() error handling (<a
href="https://redirect.github.com/eslint/eslint/issues/20734">#20734</a>)
(kuldeep kumar)</li>
<li><a
href="https://github.com/eslint/eslint/commit/4f34b1e592b0f63d766d9903998e8e36eb49d3aa"><code>4f34b1e</code></a>
chore: update pnpm/action-setup action to v5 (<a
href="https://redirect.github.com/eslint/eslint/issues/20762">#20762</a>)
(renovate[bot])</li>
<li><a
href="https://github.com/eslint/eslint/commit/51080eb5c98d619434e4835dbe9f1c6654aca3b8"><code>51080eb</code></a>
test: processor service (<a
href="https://redirect.github.com/eslint/eslint/issues/20731">#20731</a>)
(kuldeep kumar)</li>
<li><a
href="https://github.com/eslint/eslint/commit/e7e1889fca9b6044e08f41b38df20a1ce45808c8"><code>e7e1889</code></a>
chore: remove stale babel-eslint10 fixture and test (<a
href="https://redirect.github.com/eslint/eslint/issues/20727">#20727</a>)
(kuldeep kumar)</li>
<li><a
href="https://github.com/eslint/eslint/commit/4e1a87cb8fb90e309524bc36bc5f31b9f9cfaa76"><code>4e1a87c</code></a>
test: remove redundant async/await in flat config array tests (<a
href="https://redirect.github.com/eslint/eslint/issues/20722">#20722</a>)
(Pixel998)</li>
<li><a
href="https://github.com/eslint/eslint/commit/066eabb3643b12931f991594969bcc0028f71a5f"><code>066eabb</code></a>
test: add rule metadata coverage for <code>languages</code> and
<code>docs.dialects</code> (<a
href="https://redirect.github.com/eslint/eslint/issues/20717">#20717</a>)
(Pixel998)</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/eslint/eslint/commit/4d1d8f9737236603f64bbe83d5bb8001627b5611"><code>4d1d8f9</code></a>
10.2.1</li>
<li><a
href="https://github.com/eslint/eslint/commit/3e33105b05d09b5a4eb894ed75a9811fb40d65e6"><code>3e33105</code></a>
Build: changelog update for 10.2.1</li>
<li><a
href="https://github.com/eslint/eslint/commit/ca92ca0fb4599e8de1e2fb914e695fe7397cbe63"><code>ca92ca0</code></a>
docs: reuse markdown-it instance for markdown filter (<a
href="https://redirect.github.com/eslint/eslint/issues/20768">#20768</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/7ddfea9c4f62add1588c5c0b0da568c299246383"><code>7ddfea9</code></a>
chore: update dependency prettier to v3.8.2 (<a
href="https://redirect.github.com/eslint/eslint/issues/20770">#20770</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/57d2ee213305cee0cb55ef08e0480b57396269a9"><code>57d2ee2</code></a>
docs: Enable Eleventy incremental mode for watch (<a
href="https://redirect.github.com/eslint/eslint/issues/20767">#20767</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/c1621b915742276e5f4b25efe790ca62296330dc"><code>c1621b9</code></a>
docs: fix typos in code-path-analyzer.js (<a
href="https://redirect.github.com/eslint/eslint/issues/20700">#20700</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/fac40e1de2ba7646cc7cd2d3f93fbdd1f8819001"><code>fac40e1</code></a>
ci: bump pnpm/action-setup from 5.0.0 to 6.0.0 (<a
href="https://redirect.github.com/eslint/eslint/issues/20763">#20763</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/7246f923332522d8b3d46b6ee646fce88535f3fb"><code>7246f92</code></a>
test: add tests for SuppressionsService.load() error handling (<a
href="https://redirect.github.com/eslint/eslint/issues/20734">#20734</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/4f34b1e592b0f63d766d9903998e8e36eb49d3aa"><code>4f34b1e</code></a>
chore: update pnpm/action-setup action to v5 (<a
href="https://redirect.github.com/eslint/eslint/issues/20762">#20762</a>)</li>
<li><a
href="https://github.com/eslint/eslint/commit/1418d522d10bde1960f4942afb548bc7160ec49e"><code>1418d52</code></a>
docs: Update README</li>
<li>Additional commits viewable in <a
href="https://github.com/eslint/eslint/compare/v10.2.0...v10.2.1">compare
view</a></li>
</ul>
</details>
<br />


Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore <dependency name> major version` will close this
group update PR and stop Dependabot creating any more for the specific
dependency's major version (unless you unignore this specific
dependency's major version or upgrade to it yourself)
- `@dependabot ignore <dependency name> minor version` will close this
group update PR and stop Dependabot creating any more for the specific
dependency's minor version (unless you unignore this specific
dependency's minor version or upgrade to it yourself)
- `@dependabot ignore <dependency name>` will close this group update PR
and stop Dependabot creating any more for the specific dependency
(unless you unignore this specific dependency or upgrade to it yourself)
- `@dependabot unignore <dependency name>` will remove all of the ignore
conditions of the specified dependency
- `@dependabot unignore <dependency name> <ignore condition>` will
remove the ignore condition of the specified dependency and ignore
conditions


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [packaging](https://github.com/pypa/packaging) from 26.0 to 26.1.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/pypa/packaging/releases">packaging's
releases</a>.</em></p>
<blockquote>
<h2>26.1</h2>
<p>Features:</p>
<ul>
<li><del>PEP 783: add handling for Emscripten wheel tags by <a
href="https://github.com/hoodmane"><code>@​hoodmane</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/804">pypa/packaging#804</a></del>
(old name used in implementation, will be fixed in next release)</li>
<li>PEP 803: add handling for the <code>abi3.abi3t</code> free-threading
tag by <a
href="https://github.com/ngoldbaum"><code>@​ngoldbaum</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1099">pypa/packaging#1099</a></li>
<li>PEP 723: add <code>packaging.dependency_groups</code> module, based
on the <code>dependency-groups</code> package by <a
href="https://github.com/sirosen"><code>@​sirosen</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1065">pypa/packaging#1065</a></li>
<li>Add the <code>packaging.direct_url</code> module by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/944">pypa/packaging#944</a></li>
<li>Add the <code>packaging.errors</code> module by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1071">pypa/packaging#1071</a></li>
<li>Add <code>SpecifierSet.is_unsatisfiable</code> using ranges (new
internals that will be expanded in future versions) by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1119">pypa/packaging#1119</a></li>
<li>Add <code>create_compatible_tags_selector</code> to select
compatible tags by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1110">pypa/packaging#1110</a></li>
<li>Add a <code>key</code> argument to
<code>SpecifierSet.filter()</code> by <a
href="https://github.com/frostming"><code>@​frostming</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1068">pypa/packaging#1068</a></li>
<li>Support <code>&amp;</code> and <code>|</code> for
<code>Marker</code>'s by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1146">pypa/packaging#1146</a></li>
<li>Normalize <code>Version.__replace__</code> and add
<code>Version.from_parts</code> by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1078">pypa/packaging#1078</a></li>
<li>Add an option to validate compressed tag set sort order in
<code>parse_wheel_filename</code> by <a
href="https://github.com/r266-tech"><code>@​r266-tech</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1150">pypa/packaging#1150</a></li>
</ul>
<p>Behavior adaptations:</p>
<ul>
<li>Narrow exclusion of pre-releases for <code>&lt;V.postN</code> to
match spec by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1140">pypa/packaging#1140</a></li>
<li>Narrow exclusion of post-releases for <code>&gt;V</code> to match
spec by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1141">pypa/packaging#1141</a></li>
<li>Rename <code>format_full_version</code> to
<code>_format_full_version</code> to make it visibly private by <a
href="https://github.com/r266-tech"><code>@​r266-tech</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1125">pypa/packaging#1125</a></li>
<li>Restrict local version to ASCII by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1102">pypa/packaging#1102</a></li>
</ul>
<p>Pylock (PEP 751) updates:</p>
<ul>
<li>Add pylock <code>select</code> function by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1092">pypa/packaging#1092</a></li>
<li>Document pylock <code>select()</code> method and
<code>PylockSelectError</code> by <a
href="https://github.com/r266-tech"><code>@​r266-tech</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1153">pypa/packaging#1153</a></li>
<li>Add <code>filename</code> property to <code>PackageSdist</code> and
<code>PackageWheel</code>, more validation by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1095">pypa/packaging#1095</a></li>
<li>Give preference to path over url by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1128">pypa/packaging#1128</a></li>
<li>Validate name/version consistency in file names by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1114">pypa/packaging#1114</a></li>
</ul>
<p>Fixes:</p>
<ul>
<li>Fix <code>&gt;</code> comparison for versions with dev+local
segments by <a
href="https://github.com/veeceey"><code>@​veeceey</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1097">pypa/packaging#1097</a></li>
<li>Fix incorrect self-comparison for <code>InfinityType</code> and
<code>NegativeInfinityType</code> by <a
href="https://github.com/bysiber"><code>@​bysiber</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1093">pypa/packaging#1093</a></li>
<li>Canonicalize when deduplicating specifiers in
<code>SpecifierSet</code> by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1109">pypa/packaging#1109</a></li>
<li>Fix charset error message formatting by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1121">pypa/packaging#1121</a></li>
<li>Handle the <code>key</code> parameter in
<code>SpecifierSet.filter</code> when specifiers are empty and
prerelease is <code>False</code> by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1096">pypa/packaging#1096</a></li>
<li>Standardize inner components of <code>repr</code> output by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1090">pypa/packaging#1090</a></li>
<li><code>Specifier</code>'s <code>===</code> uses original string, not
normalized, when available by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1124">pypa/packaging#1124</a></li>
<li>Propagate int-max-str-digits <code>ValueError</code> by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1155">pypa/packaging#1155</a></li>
</ul>
<p>Performance:</p>
<ul>
<li>Add fast path for parsing simple versions (digits and dots only) by
<a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1082">pypa/packaging#1082</a></li>
<li>Add fast path for <code>Version</code> to <code>Version</code>
comparison by skipping <code>_key</code> property by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1083">pypa/packaging#1083</a></li>
<li>Cache <code>Version</code> hash value in dedicated slot by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1118">pypa/packaging#1118</a></li>
<li>Overhaul <code>_cmpkey</code> to remove use of custom objects by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1116">pypa/packaging#1116</a></li>
<li>Skip <code>__replace__</code> in Specifier comparison if not needed
by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1081">pypa/packaging#1081</a></li>
<li><code>SpecifierSet</code> use <code>tuple</code> instead of
<code>frozenset</code> for <code>_specs</code> by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1108">pypa/packaging#1108</a></li>
<li>Speed up complex <code>SpecifierSet</code> filtering by implementing
cost-based ordering by <a
href="https://github.com/notatallshaw"><code>@​notatallshaw</code></a>
in <a
href="https://redirect.github.com/pypa/packaging/pull/1105">pypa/packaging#1105</a></li>
</ul>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/pypa/packaging/blob/main/CHANGELOG.rst">packaging's
changelog</a>.</em></p>
<blockquote>
<p>26.1 - 2026-04-14</p>
<pre><code>
Features:
<ul>
<li>PEP 783: add handling for Emscripten wheel tags in
(:pull:<code>804</code>)</li>
<li>PEP 803: add handling for the <code>abi3.abi3t</code> free-threading
tag in (:pull:<code>1099</code>)</li>
<li>PEP 723: add <code>packaging.dependency_groups</code> module, based
on the <code>dependency-groups</code> package in
(:pull:<code>1065</code>)</li>
<li>Add the <code>packaging.direct_url</code> module in
(:pull:<code>944</code>)</li>
<li>Add the <code>packaging.errors</code> module in
(:pull:<code>1071</code>)</li>
<li>Add <code>SpecifierSet.is_unsatisfiable</code> using ranges (new
internals that will be expanded in future versions) in
(:pull:<code>1119</code>)</li>
<li>Add <code>create_compatible_tags_selector</code> to select
compatible tags in (:pull:<code>1110</code>)</li>
<li>Add a <code>key</code> argument to
<code>SpecifierSet.filter()</code> in (:pull:<code>1068</code>)</li>
<li>Support <code>&amp;amp;</code> and <code>|</code> for
<code>Marker</code>'s in (:pull:<code>1146</code>)</li>
<li>Normalize <code>Version.__replace__</code> and add
<code>Version.from_parts</code> in (:pull:<code>1078</code>)</li>
<li>Add an option to validate compressed tag set sort order in
<code>parse_wheel_filename</code> in (:pull:<code>1150</code>)</li>
</ul>
<p>Behavior adaptations:</p>
<ul>
<li>Narrow exclusion of pre-releases for <code>&amp;lt;V.postN</code> to
match spec in (:pull:<code>1140</code>)</li>
<li>Narrow exclusion of post-releases for <code>&amp;gt;V</code> to
match spec in (:pull:<code>1141</code>)</li>
<li>Rename <code>format_full_version</code> to
<code>_format_full_version</code> to make it visibly private in
(:pull:<code>1125</code>)</li>
<li>Restrict local version to ASCII in (:pull:<code>1102</code>)</li>
</ul>
<p>Pylock (PEP 751) updates:</p>
<ul>
<li>Add pylock <code>select</code> function in
(:pull:<code>1092</code>)</li>
<li>Document pylock <code>select()</code> method and
<code>PylockSelectError</code> in (:pull:<code>1153</code>)</li>
<li>Add <code>filename</code> property to <code>PackageSdist</code> and
<code>PackageWheel</code>, more validation in
(:pull:<code>1095</code>)</li>
<li>Give preference to path over url in (:pull:<code>1128</code>)</li>
<li>Validate name/version consistency in file names in
(:pull:<code>1114</code>)</li>
</ul>
<p>Fixes:</p>
<ul>
<li>Fix <code>&amp;gt;</code> comparison for versions with dev+local
segments in (:pull:<code>1097</code>)</li>
<li>Fix incorrect self-comparison for <code>InfinityType</code> and
<code>NegativeInfinityType</code> in (:pull:<code>1093</code>)</li>
<li>Canonicalize when deduplicating specifiers in
<code>SpecifierSet</code> in (:pull:<code>1109</code>)</li>
<li>Fix charset error message formatting in
(:pull:<code>1121</code>)</li>
<li>Handle the <code>key</code> parameter in
<code>SpecifierSet.filter</code> when specifiers are empty and
prerelease is <code>False</code> in (:pull:<code>1096</code>)</li>
<li>Standardize inner components of <code>repr</code> output in
(:pull:<code>1090</code>)</li>
<li><code>Specifier</code>'s <code>===</code> uses original string, not
normalized, when available in (:pull:<code>1124</code>)</li>
<li>Propagate int-max-str-digits <code>ValueError</code> in
(:pull:<code>1155</code>)</li>
</ul>
<p>Performance:</p>
<ul>
<li>Add fast path for parsing simple versions (digits and dots only) in
(:pull:<code>1082</code>)</li>
<li>Add fast path for <code>Version</code> to <code>Version</code>
comparison by skipping <code>_key</code> property in
(:pull:<code>1083</code>)</li>
<li>Cache <code>Version</code> hash value in dedicated slot in
(:pull:<code>1118</code>)</li>
<li>Overhaul <code>_cmpkey</code> to remove use of custom objects in
(:pull:<code>1116</code>)</li>
<li>Skip <code>__replace__</code> in Specifier comparison if not needed
in (:pull:<code>1081</code>)<br />
&lt;/tr&gt;&lt;/table&gt;<br />
</code></pre></li>
</ul>
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/pypa/packaging/commit/c1a88a3e035e8bfe47dbc957f4a2493e8a7b4f3c"><code>c1a88a3</code></a>
Bump for release</li>
<li><a
href="https://github.com/pypa/packaging/commit/702c25ecc297054f9597acda987a75dd6833acf5"><code>702c25e</code></a>
docs: update changelog for 26.1 (<a
href="https://redirect.github.com/pypa/packaging/issues/1156">#1156</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/3f4f5d46808f51709aba8341d4121f20752b3235"><code>3f4f5d4</code></a>
Implement <code>is_unsatisfiable</code> on <code>SpecifierSet</code>
using ranges (<a
href="https://redirect.github.com/pypa/packaging/issues/1119">#1119</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/06c6555f44f0f7b84459c876375c5b1d52ee7fef"><code>06c6555</code></a>
Propagate int-max-str-digits ValueError (<a
href="https://redirect.github.com/pypa/packaging/issues/1155">#1155</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/905c90c1eb8c77dede11899194ca4b0f5eaf188d"><code>905c90c</code></a>
feat: option to validate compressed tag set sort order in
`parse_wheel_filena...</li>
<li><a
href="https://github.com/pypa/packaging/commit/af0026cff97a8f28d165d6ac4afb58b2cdc8ffc5"><code>af0026c</code></a>
docs(pylock): document select() method and PylockSelectError (<a
href="https://redirect.github.com/pypa/packaging/issues/1153">#1153</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/668da8662a112036e4850b68980790cb2d10dae9"><code>668da86</code></a>
Rename format_full_version to _format_full_version to make it visibly
private...</li>
<li><a
href="https://github.com/pypa/packaging/commit/f294d529e2634798a98fd9cee6708307a5c1792c"><code>f294d52</code></a>
tests: do not reload the tags module (<a
href="https://redirect.github.com/pypa/packaging/issues/1152">#1152</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/2c6c7dfa2d406e44b91f3f8d26eeed72437c9435"><code>2c6c7df</code></a>
feat: add handling for Emscripten wheels tags per PEP 783 (<a
href="https://redirect.github.com/pypa/packaging/issues/804">#804</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/6762eea4de40da02c119d02cc77c907c6b446b86"><code>6762eea</code></a>
docs(markers): document &amp; and | operators for combining Marker
objects (<a
href="https://redirect.github.com/pypa/packaging/issues/1151">#1151</a>)</li>
<li>Additional commits viewable in <a
href="https://github.com/pypa/packaging/compare/26.0...26.1">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=packaging&package-manager=pip&previous-version=26.0&new-version=26.1)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…inotoolkit#34669)

This fixes issue openvinotoolkit#34668

Co-authored-by: Pavel Durandin <pavel.durandin@intel.com>
… in the npm-packages group (openvinotoolkit#35439)

Bumps the npm-packages group in /src/bindings/js/node with 1 update:
[typescript-eslint](https://github.com/typescript-eslint/typescript-eslint/tree/HEAD/packages/typescript-eslint).

Updates `typescript-eslint` from 8.58.2 to 8.59.0
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/typescript-eslint/typescript-eslint/releases">typescript-eslint's
releases</a>.</em></p>
<blockquote>
<h2>v8.59.0</h2>
<h2>8.59.0 (2026-04-20)</h2>
<h3>🚀 Features</h3>
<ul>
<li><strong>eslint-plugin:</strong> [no-unnecessary-type-assertion]
report more cases based on assignability (<a
href="https://redirect.github.com/typescript-eslint/typescript-eslint/pull/11789">#11789</a>)</li>
</ul>
<h3>❤️ Thank You</h3>
<ul>
<li>Ulrich Stark</li>
</ul>
<p>See <a
href="https://github.com/typescript-eslint/typescript-eslint/releases/tag/v8.59.0">GitHub
Releases</a> for more information.</p>
<p>You can read about our <a
href="https://typescript-eslint.io/users/versioning">versioning
strategy</a> and <a
href="https://typescript-eslint.io/users/releases">releases</a> on our
website.</p>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/typescript-eslint/typescript-eslint/blob/main/packages/typescript-eslint/CHANGELOG.md">typescript-eslint's
changelog</a>.</em></p>
<blockquote>
<h2>8.59.0 (2026-04-20)</h2>
<p>This was a version bump only for typescript-eslint to align it with
other projects, there were no code changes.</p>
<p>See <a
href="https://github.com/typescript-eslint/typescript-eslint/releases/tag/v8.59.0">GitHub
Releases</a> for more information.</p>
<p>You can read about our <a
href="https://typescript-eslint.io/users/versioning">versioning
strategy</a> and <a
href="https://typescript-eslint.io/users/releases">releases</a> on our
website.</p>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/typescript-eslint/typescript-eslint/commit/ea9ae4f8817873480e3501145059f63e39e8d8a1"><code>ea9ae4f</code></a>
chore(release): publish 8.59.0</li>
<li>See full diff in <a
href="https://github.com/typescript-eslint/typescript-eslint/commits/v8.59.0/packages/typescript-eslint">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=typescript-eslint&package-manager=npm_and_yarn&previous-version=8.58.2&new-version=8.59.0)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore <dependency name> major version` will close this
group update PR and stop Dependabot creating any more for the specific
dependency's major version (unless you unignore this specific
dependency's major version or upgrade to it yourself)
- `@dependabot ignore <dependency name> minor version` will close this
group update PR and stop Dependabot creating any more for the specific
dependency's minor version (unless you unignore this specific
dependency's minor version or upgrade to it yourself)
- `@dependabot ignore <dependency name>` will close this group update PR
and stop Dependabot creating any more for the specific dependency
(unless you unignore this specific dependency or upgrade to it yourself)
- `@dependabot unignore <dependency name>` will remove all of the ignore
conditions of the specified dependency
- `@dependabot unignore <dependency name> <ignore condition>` will
remove the ignore condition of the specified dependency and ignore
conditions


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
jkasprza and others added 30 commits April 24, 2026 07:52
### Details:
- Fix L0 build issues introduced by
openvinotoolkit#34974
- Replace direct calls to `dnnl::ocl_interop` with
`get_onednn_grouped_memory` and `get_onednn_memory` from OpenVINO memory
API.
### AI Assistance:
 - *AI assistance used: no
### Details:
- Remove `/wd4244` and `/wd4018` global warning suppression from GPU
plugin `CMakeLists.txt` and fix remaining C4244 narrowing conversion
warnings for BinSkim BA2007 compliance (final step)
- **CMake:** Remove `/wd4244` `/wd4018` from plugin `CMakeLists.txt`;
relocate to `tests/CMakeLists.txt` (test binaries are not subject to
BinSkim). Remove duplicate `/wd4244` from
`tests/functional/CMakeLists.txt`.
- **prior_box.cpp:** Widen `int` → `int64_t` for loop vars/accumulators;
`sqrt` → `sqrtf`; `static_cast<float>` for `std::round` results
- **paged_attention_opt.cpp:** `auto` → `int64_t` for `aligned_seq_len`
and `seq_length`
- **variable_state.cpp:** Remove erroneous `0.f` third arg from
`cldnn::padding` ctor (parameter is `DynamicDimsMask`, not `float`)
- **plugin.cpp:** Lambda param `float` → `double`; cast
`max_global_cache_size` to `float` at declaration
- **utils.cpp:** `static_cast<int>(inner_blks[i])` at oneDNN API
boundary

### Tickets:
- [CVS-185007](https://jira.devtools.intel.com/browse/CVS-185007)

### AI Assistance:
- AI assistance used: yes
- GitHub Copilot used for code analysis (identifying root causes of each
warning, evaluating fix options for type safety), generating fix code,
and verifying build results with `CMAKE_COMPILE_WARNING_AS_ERROR=ON`
(`/WX`). All changes were human-reviewed for correctness, validated via
local `/WX` clean build with `ENABLE_TESTS=ON`, and confirmed via
BinSkim BA2007 scan to have zero GPU plugin violations.
…it#35337)

### Details:
 - Add Mixture of Experts support to model builder

### Tickets:
 - *EISW-209517*

### AI Assistance:
 - *AI assistance used: yes*
- *Claude Code used for coding assistance and to analyze GPT OSS model
to recreate patterns, manually tested and verified*

Correct subgraphs produced

Generate:
```
{
  "pipeline_name": "npuw_synthetic_decoder_kv1152",
  "repeated_numbers": {
    "total_subgraphs": 1,
    "synthetic_decoder_kv1152_0_FCEW000": 1
  },
  "subgraphs": [
    "synthetic_decoder_kv1152_0_FCEW000.xml"
  ]
}
```
Prefill:
```
{
  "pipeline_name": "npuw_synthetic_decoder_prefill",
  "repeated_numbers": {
    "total_subgraphs": 5,
    "synthetic_decoder_prefill_00_FCEW000": 1,
    "synthetic_decoder_prefill_01_REP009C": 12,
    "synthetic_decoder_prefill_02_REP00B7": 12,
    "synthetic_decoder_prefill_03_REP00B9": 11,
    "synthetic_decoder_prefill_36_FCEW001": 1
  },
  "subgraphs": [
    "synthetic_decoder_prefill_00_FCEW000.xml",
    "synthetic_decoder_prefill_01_REP009C.xml",
    "synthetic_decoder_prefill_03_REP00B9.xml",
    "synthetic_decoder_prefill_36_FCEW001.xml"
  ],
  "moe_subgraphs": [
    "synthetic_decoder_prefill_02_REP00B7_moe_chunk_16.xml",
    "synthetic_decoder_prefill_02_REP00B7_moe_chunk_32.xml",
    "synthetic_decoder_prefill_02_REP00B7_moe_chunk_64.xml",
    "synthetic_decoder_prefill_02_REP00B7_moe_chunk_128.xml",
    "synthetic_decoder_prefill_02_REP00B7_moe_chunk_256.xml"
  ]

```
Bumps [certifi](https://github.com/certifi/python-certifi) from
2026.2.25 to 2026.4.22.
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/certifi/python-certifi/commit/5dddfb072243da27adde885b73ba9b809c3224ca"><code>5dddfb0</code></a>
2026.04.22 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/410">#410</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/f99eccdaf87f7c10e521a58a700ca3eb94a0787e"><code>f99eccd</code></a>
Bump peter-evans/create-pull-request from 8.1.0 to 8.1.1 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/404">#404</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/918bed055f7291719512af186c1c24710f845660"><code>918bed0</code></a>
Bump actions/upload-artifact from 7.0.0 to 7.0.1 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/405">#405</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/0a49067eb434e53e1f8df5f7707d5dc05ef9def4"><code>0a49067</code></a>
Bump pypa/gh-action-pypi-publish from 1.13.0 to 1.14.0 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/403">#403</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/acf6ce8e39e3b125f4349e11904295e4fe4c1bed"><code>acf6ce8</code></a>
Bump actions/download-artifact from 8.0.0 to 8.0.1 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/398">#398</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/feb0ed26163a9417ea0fb8eb52d47e79fcf202ab"><code>feb0ed2</code></a>
Bump actions/download-artifact from 7.0.0 to 8.0.0 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/397">#397</a>)</li>
<li><a
href="https://github.com/certifi/python-certifi/commit/d9c11a50369cc377abb40f7909ded3d6da4d98a3"><code>d9c11a5</code></a>
Bump actions/upload-artifact from 6.0.0 to 7.0.0 (<a
href="https://redirect.github.com/certifi/python-certifi/issues/396">#396</a>)</li>
<li>See full diff in <a
href="https://github.com/certifi/python-certifi/compare/2026.02.25...2026.04.22">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=certifi&package-manager=pip&previous-version=2026.2.25&new-version=2026.4.22)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [idna](https://github.com/kjd/idna) from 3.11 to 3.13.
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/kjd/idna/blob/master/HISTORY.rst">idna's
changelog</a>.</em></p>
<blockquote>
<p>3.13 (2026-04-22)
+++++++++++++++++</p>
<ul>
<li>Correct classification error for codepoint U+A7F1</li>
</ul>
<p>3.12 (2026-04-21)
+++++++++++++++++</p>
<ul>
<li>Update to Unicode 17.0.0.</li>
<li>Issue a deprecation warning for the transitional argument.</li>
<li>Added lazy-loading to provide some performance improvements.</li>
<li>Removed vestiges of code related to Python 2 support, including
segmentation of data structures specific to Jython.</li>
</ul>
<p>Thanks to Rodrigo Nogueira for contributions to this release.</p>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/kjd/idna/commit/89cdfd27338896cee6b1ee18e64c96ac28684ce0"><code>89cdfd2</code></a>
Release v3.13</li>
<li><a
href="https://github.com/kjd/idna/commit/1eb068687543118147417a8d8a70674e2c172891"><code>1eb0686</code></a>
Pre-release 3.13</li>
<li><a
href="https://github.com/kjd/idna/commit/5f20d1e41eea3b3873d18d83d7a384784f72a92e"><code>5f20d1e</code></a>
Merge pull request <a
href="https://redirect.github.com/kjd/idna/issues/220">#220</a> from
kjd/unicode-next</li>
<li><a
href="https://github.com/kjd/idna/commit/4ea84252ab21e62a79e5a3273746112b5dcfb810"><code>4ea8425</code></a>
Regenerate idnadata.py with correct NFKC_CF data</li>
<li><a
href="https://github.com/kjd/idna/commit/fd47341a08bbdcffda33694211ca4de10170cd41"><code>fd47341</code></a>
Use NFKC_CF from Unicode data files instead of Python's unicodedata
module</li>
<li><a
href="https://github.com/kjd/idna/commit/a5304a4cdbd7b31595f8ac42ffdfa88f5b936467"><code>a5304a4</code></a>
Merge pull request <a
href="https://redirect.github.com/kjd/idna/issues/219">#219</a> from
kjd/release-3.12</li>
<li><a
href="https://github.com/kjd/idna/commit/d80d6f9254d699961fa2c669a1534cde9d4ee5b6"><code>d80d6f9</code></a>
Release v3.12</li>
<li><a
href="https://github.com/kjd/idna/commit/1bb44ddb3f2a9dcf97a6ac11aba34e5b6ed31291"><code>1bb44dd</code></a>
Merge pull request <a
href="https://redirect.github.com/kjd/idna/issues/218">#218</a> from
kjd/release-candidate-3.12rc0</li>
<li><a
href="https://github.com/kjd/idna/commit/909c49d15b8d159be163bccc7972116baffdb47b"><code>909c49d</code></a>
Release candidate for 3.12</li>
<li><a
href="https://github.com/kjd/idna/commit/c5459a10370f005dc09921aee3201b5a45699f9d"><code>c5459a1</code></a>
Merge pull request <a
href="https://redirect.github.com/kjd/idna/issues/217">#217</a> from
kjd/housekeeping-2</li>
<li>Additional commits viewable in <a
href="https://github.com/kjd/idna/compare/v3.11...v3.13">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=idna&package-manager=pip&previous-version=3.11&new-version=3.13)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…gs/python (openvinotoolkit#35509)

Updates the requirements on
[pyright](https://github.com/RobertCraigie/pyright-python) to permit the
latest version.
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/d7508e5425d3d02761d70dab1f9a2086573af429"><code>d7508e5</code></a>
[pyright updated to 1.1.409] Update Version (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/359">#359</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/81b795a41ddcc3c77218d8c8e406983e39852285"><code>81b795a</code></a>
Pyright NPM Package update to 1.1.408 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/357">#357</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/53e8efb4638daca6c541d18ad35dafb0e5d9f34d"><code>53e8efb</code></a>
Pyright NPM Package update to 1.1.407 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/356">#356</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/1d515b71299f9dc4c066ce37e673d93bda4bc2df"><code>1d515b7</code></a>
Pyright NPM Package update to 1.1.406 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/355">#355</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/e211ec8df8d389a147af6d31bdb2bc812f04aa5a"><code>e211ec8</code></a>
Pyright NPM Package update to 1.1.405 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/353">#353</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/d393df1703a808473b84bd14a2702f4793014031"><code>d393df1</code></a>
Pyright NPM Package update to 1.1.404 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/352">#352</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/047488f6b38c6bd59e43daf838b2863b6cab9008"><code>047488f</code></a>
Pyright NPM Package update to 1.1.403 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/350">#350</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/708a9d4a964376a7be931b7721d315ef1d2db31d"><code>708a9d4</code></a>
Pyright NPM Package update to 1.1.402 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/349">#349</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/8df87d18e0baad8b65afd8488bf27a90df5cdcb3"><code>8df87d1</code></a>
Pyright NPM Package update to 1.1.401 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/347">#347</a>)</li>
<li><a
href="https://github.com/RobertCraigie/pyright-python/commit/7e1526dc0349fedbae25b3b2d0d105ca55a2a8fb"><code>7e1526d</code></a>
Pyright NPM Package update to 1.1.400 (<a
href="https://redirect.github.com/RobertCraigie/pyright-python/issues/346">#346</a>)</li>
<li>Additional commits viewable in <a
href="https://github.com/RobertCraigie/pyright-python/compare/v0.0.1...v1.1.409">compare
view</a></li>
</ul>
</details>
<br />


Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
### Details:
 - *Base class for PagedDeltaNet*
 - *...*

### Tickets:
 - *CVS-183795*

### AI Assistance:
 - *AI assistance used: yes*
 - *Generate op draft based on spec and similar existing internal ops.*

---------

Co-authored-by: Copilot <copilot@github.com>
### Details:
Fix nightly coverage build failure: scheduled runs were overriding the
pinned onednn_gpu
submodule with oneDNN main, which is incompatible with the current
OpenVINO GPU code and
caused the XE3P build error.

**Change**: set use_latest_onednn: false in coverage.yml so nightly
coverage builds use the
pinned oneDNN revision, same as regular CI builds.

### AI Assistance:
 - *AI assistance used: yes*
 - *Analyzing multiple build logs to identify root cause.*
### Details:
 - *Base class for PagedCausalConv1d*
 - *...*

### Tickets:
 - *CVS-183794*

### AI Assistance:
 - *AI assistance used: yes*
 - *Generate op draft based on spec and similar existing internal ops.*

---------

Co-authored-by: Copilot <copilot@github.com>
…vinotoolkit#35487)

### Details:
- removed hardcoded implementations to skip PagedAttentionExtension
inputs

### Tickets:
[CVS-183796](https://jira.devtools.intel.com/browse/CVS-183796)
[CVS-183797](https://jira.devtools.intel.com/browse/CVS-183797)

### AI Assistance:
 - *AI assistance used: no*

---------

Co-authored-by: Copilot <copilot@github.com>
…vinotoolkit#35410)

### Details:
- Migration proposal for ov::Dimension ctor to use string view instead
of string
- Migrate common utils to use string view whre possible. The `join` util
skipped as it will be done separately.
 - Remove not used utils
 - Expand tests
- Update code to use new utils version, in some place explicit convert
to string from view is required.

### Tickets:
 - N/A

### AI Assistance:
 - *AI assistance used: yes*
 - *generate doxy, refactor code, add tests*

---------

Signed-off-by: Pawel Raasz <pawel.raasz@intel.com>
Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
### Details:
 - removing metadata padding since it is no longer needed

### Tickets:
 - *N/A*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: alexandruenache1111 <alexandru.enache@intel.com>
…olkit#35240)

### Details:
- Update the NPUW test model builder’s RoPE frequency construction to
better match the RopeCache/RopePatternLLama2 graph pattern .

### Tickets:
 - EISW-209517

### AI Assistance:
 - *yes*
 - *Claude code used to find the issue and implement fix*
…t#35500)

### Details:


### Tickets:
- [CVS-183796](https://jira.devtools.intel.com/browse/CVS-183796)
- [CVS-183797](https://jira.devtools.intel.com/browse/CVS-183797)

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Co-authored-by: Copilot <copilot@github.com>
Bumps [packaging](https://github.com/pypa/packaging) from 26.1 to 26.2.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/pypa/packaging/releases">packaging's
releases</a>.</em></p>
<blockquote>
<h2>26.2</h2>
<h2>What's Changed</h2>
<p>Fixes:</p>
<ul>
<li>Fix incorrect sysconfig var name for pyemscripten by <a
href="https://github.com/ryanking13"><code>@​ryanking13</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1160">pypa/packaging#1160</a></li>
<li>Make <code>Version</code>, <code>Specifier</code>,
<code>SpecifierSet</code>, <code>Tag</code>, <code>Marker</code>, and
<code>Requirement</code> pickle-safe
and backward-compatible with pickles created in 25.0-26.1 (including
references to the removed
<code>packaging._structures</code> module) by <a
href="https://github.com/eachimei"><code>@​eachimei</code></a> and <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1163">pypa/packaging#1163</a>,
<a
href="https://redirect.github.com/pypa/packaging/pull/1168">pypa/packaging#1168</a>,
<a
href="https://redirect.github.com/pypa/packaging/pull/1170">pypa/packaging#1170</a>,
and <a
href="https://redirect.github.com/pypa/packaging/pull/1171">pypa/packaging#1171</a></li>
<li>fix: re-export ExceptionGroup for now by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1164">pypa/packaging#1164</a></li>
</ul>
<p>Documentation:</p>
<ul>
<li>docs: add errors section and fix missing details by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1159">pypa/packaging#1159</a></li>
<li>docs(dev): document property-based test suite by <a
href="https://github.com/r266-tech"><code>@​r266-tech</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1167">pypa/packaging#1167</a></li>
<li>Fix typo in DirectUrl documentation by <a
href="https://github.com/sbidoul"><code>@​sbidoul</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1169">pypa/packaging#1169</a></li>
<li>docs(specifiers): add is_unsatisfiable() usage example by <a
href="https://github.com/r266-tech"><code>@​r266-tech</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1166">pypa/packaging#1166</a></li>
</ul>
<p>Internal:</p>
<ul>
<li>Enable the auditor persona on zizmor by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1158">pypa/packaging#1158</a></li>
<li>Test new pickle guarantees by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1174">pypa/packaging#1174</a></li>
<li>Use native uv integration in rtd by <a
href="https://github.com/henryiii"><code>@​henryiii</code></a> in <a
href="https://redirect.github.com/pypa/packaging/pull/1175">pypa/packaging#1175</a></li>
</ul>
<h2>New Contributors</h2>
<ul>
<li><a
href="https://github.com/ryanking13"><code>@​ryanking13</code></a> made
their first contribution in <a
href="https://redirect.github.com/pypa/packaging/pull/1160">pypa/packaging#1160</a></li>
<li><a href="https://github.com/eachimei"><code>@​eachimei</code></a>
made their first contribution in <a
href="https://redirect.github.com/pypa/packaging/pull/1163">pypa/packaging#1163</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/pypa/packaging/compare/26.1...26.2">https://github.com/pypa/packaging/compare/26.1...26.2</a></p>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/pypa/packaging/blob/main/CHANGELOG.rst">packaging's
changelog</a>.</em></p>
<blockquote>
<p>26.2 - 2026-04-24</p>
<pre><code>
Fixes:
<ul>
<li>Fix incorrect sysconfig var name for pyemscripten in
(:pull:<code>1160</code>)</li>
<li>Make <code>Version</code>, <code>Specifier</code>,
<code>SpecifierSet</code>, <code>Tag</code>, <code>Marker</code>, and
<code>Requirement</code> pickle-safe<br />
and backward-compatible with pickles created in 25.0-26.1 (including
references to the removed<br />
<code>packaging._structures</code> module) (:pull:<code>1163</code>,
:pull:<code>1168</code>, :pull:<code>1170</code>,
:pull:<code>1171</code>)</li>
<li>Re-export <code>ExceptionGroup</code> in metadata for now in
(:pull:<code>1164</code>)</li>
</ul>
<p>Documentation:</p>
<ul>
<li>Add errors section and fix missing details in
(:pull:<code>1159</code>)</li>
<li>Document our property-based test suite in
(:pull:<code>1167</code>)</li>
<li>Fix a <code>DirectUrl</code> typo in (:pull:<code>1167</code>)</li>
<li>Add example of <code>is_unsatisfiable</code> in
(:pull:<code>1166</code>)</li>
</ul>
<p>Internal:</p>
<ul>
<li>Enable the auditor persona on zizmor in
(:pull:<code>1158</code>)</li>
<li>Test new pickle guarantees in (:pull:<code>1174</code>)</li>
<li>Use new native ReadTheDocs uv integration in
(:pull:<code>1175</code>)<br />
</code></pre></li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/pypa/packaging/commit/84a87ee42483d7352f9502d78a9553da8859aa7a"><code>84a87ee</code></a>
Bump for release</li>
<li><a
href="https://github.com/pypa/packaging/commit/4a616b65bed23c8c6d58e6b0fc1a4434d4ff1f14"><code>4a616b6</code></a>
docs: a few more updates to prepare for 26.2 (<a
href="https://redirect.github.com/pypa/packaging/issues/1176">#1176</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/9de6f44f1e82d4595edf3aad1c4f6f98c85935a0"><code>9de6f44</code></a>
ci: use native uv integration in rtd (<a
href="https://redirect.github.com/pypa/packaging/issues/1175">#1175</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/bc76e14debd1a2799d1ca8f9d9c9823f35bfa466"><code>bc76e14</code></a>
chore: update changelog for 26.2 (<a
href="https://redirect.github.com/pypa/packaging/issues/1161">#1161</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/3f00091c08f0aa830e33ed7db00f16f11c8ac97f"><code>3f00091</code></a>
tests: add a pickle check (<a
href="https://redirect.github.com/pypa/packaging/issues/1174">#1174</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/48a8a069805291186522de3eff73ea80a8ca96ad"><code>48a8a06</code></a>
fix: make Requirements/Markers pickle-safe (<a
href="https://redirect.github.com/pypa/packaging/issues/1171">#1171</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/823b44ed1f904084a77ae3adf0ef130af6365f84"><code>823b44e</code></a>
fix: make Tags pickle-safe (<a
href="https://redirect.github.com/pypa/packaging/issues/1170">#1170</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/4bed32d920ca7211dd65fdf0a1ee06376e9c4733"><code>4bed32d</code></a>
fix: make Specifier / SpecifierSet pickle-safe (<a
href="https://redirect.github.com/pypa/packaging/issues/1168">#1168</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/963118e37caae97bc8b72f72956c7fb4ca9857ec"><code>963118e</code></a>
fix: re-export ExceptionGroup for now (<a
href="https://redirect.github.com/pypa/packaging/issues/1164">#1164</a>)</li>
<li><a
href="https://github.com/pypa/packaging/commit/66e34a80256c96dea11da143682950c84b8133bb"><code>66e34a8</code></a>
docs(specifiers): add is_unsatisfiable() usage example (<a
href="https://redirect.github.com/pypa/packaging/issues/1166">#1166</a>)</li>
<li>Additional commits viewable in <a
href="https://github.com/pypa/packaging/compare/26.1...26.2">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=packaging&package-manager=pip&previous-version=26.1&new-version=26.2)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)


</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
**Details:** Specify PagedCausalConv1D operation

**Ticket:** CVS-183792

### AI Assistance:
 - *AI assistance used: yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
…ompiled file; employ CI Doctor for analysis of `Debian 10 ARM` failures (openvinotoolkit#35548)

### Tickets:
 - *182215*

### AI Assistance:
 - *AI assistance used: yes*
- Proofreading and suggesting boundaries for the CI Doctor agentic
workflow.
### Details:
 - 

### Tickets:
 - *E-194204*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: alexandruenache1111 <alexandru.enache@intel.com>
)

### Details:
 - Add a new OCL implementation for fsv16 group normalization
- The new implementation is used if each group contains fewer than
fsv=16 features
- A single fused kernel handles all stages of the reduction, avoiding
excessive loading of shared values and reusing cache in cases of small
inputs

### Tickets:
 - CVS-177816

---------

Co-authored-by: Roman Lyamin <Roman.Lyamin@intel.com>
…at types for state table (openvinotoolkit#35540)

### Details:
 - *Allow independent float types for state table*
- *Based on
openvinotoolkit@043c0c9
by @liubo-intel*

### Tickets:
 - *ticket-id*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*
…loat type (openvinotoolkit#35541)

### Details:
 - *type validation for independent state float type*
 - *related to openvinotoolkit#35540

### Tickets:
 - *ticket-id*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

Co-authored-by: Copilot <copilot@github.com>
…olkit#35547)

### Details:
Some dGPU configs have know limitation on the max host tensor size (see
`can_use_usm_host` implementation).
In the case of dynamic shapes, it's impossible to estimate the maximum
tensor size. This led to the fact that the host output tensor is used
for any output tensor size, and if it's large enough, we get visible
performance drop on some specific HW.
It was decided to unconditionally disable this reuse for discrete GPUs
to keep performance intact.

### Tickets:
 - CVS-184515

### AI Assistance:
 - *AI assistance used: no
… on Windows (openvinotoolkit#35498)

### Details:
Cleaning remaining stuff left after troubleshooting of `Content-Type:
Unknown` error during `pip install` on Windows - see
openvinotoolkit#34376

Workarounds are currently in testing:
- openvinotoolkit#35494
- openvinotoolkit#34376
- Updated Python API stub (.pyi) files from the latest available nightly

Auto-generated by GitHub Actions

Co-authored-by: Alicja Miloszewska <alicja.miloszewska@intel.com>
### Details:
* Added a new transformation pass `PositionIDsReplacerLFM2` that detects
and replaces arange-based position computations in LFM2-style models
with explicit `position_ids` inputs.
* Implemented comprehensive unit tests for the new LFM2 position ID
replacement
* Renamed files and updated includes to move `position_ids_replacer`
from the `sdpa_to_paged_attention` subdirectory to a new
`paged_attention` directory

### Tickets:
 - *CVS-183796*
 - *CVS-183797*

### AI Assistance:
 - *AI assistance used: yes*
 - *initial development*

Co-authored-by: Andrii Staikov <andrii.staikov@intel.com>
### Details:
 - *item1*
 - *...*

### Tickets:
 - [CVS-179287](https://jira.devtools.intel.com/browse/CVS-179287)

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*
)

### Details:
Update pass for adding attention mask for Whisper decoder model
(AttentionMaskInput) as pattern has changed after transformers update on
newer version (v5).

### AI Assistance:
 - *AI assistance used: no *
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*
### Details:
- Fixed SWA Mask generation with static shape: use the real query index.
- Code refactored: file renamed and wrapped helper
`rebuild_sliding_window_mask` to reuse among different rewriters.

### Tickets:
 - *[EISW-211588](https://jira.devtools.intel.com/browse/EISW-211588)*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: intelgaoxiong <xiong.gao@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.