[GPU] fix dGPU performance issue when run qwen3_moe models (#35135)

[NPU]Move VCL API from compiler_adapter to utils (#34662)

[NPU] Update getCompilerVersion and getCompilerProperties methods (#35152)

Minor improvements to `token_type_ids` extension for PA (#34661)

SDL512: Fix npm audit vulnerabilities in package-lock.json (#35114)

[core] Weight sharing manage context lifetime (#34845)

[NPU][WS] Use "ov::enable_weightless" instead of the "weightless_blob" NPU specific config option (#35063)

[NPU] Add support for compiled model release memory (#35084)

Eliminate gold linker usage (#35112)

Pin Keras to workaround Tensorflow Lite Layer Tests failure (#35159)

[NPUW] Model builder: fix Whisper decoder synthetic model (#35055)

[NPU] Compute hash inside CommandQueueDesc and not in the graph (#35104)

[CPU][ARM] `AclPoolingExecutor` should not use default parameter values if provided ones are incorrect (#35110)

Bump requests from 2.32.5 to 2.33.1 (#35079)

[GPU] Fix rank-changing reorder fusion for depth_to_space (#35099)