Rebase 2025.01.21

[kzawora-intel]

• [Bugfix] Fix score api for missing max_model_len validation ([Bugfix] Fix score api for missing max_model_len validation vllm-project/vllm#12119)
• [Bugfix] Mistral tokenizer encode accept list of str ([Bugfix] Mistral tokenizer encode accept list of str vllm-project/vllm#12149)
• [AMD][FP8] Using MI300 FP8 format on ROCm for block_quant ([AMD][FP8] Using MI300 FP8 format on ROCm for block_quant vllm-project/vllm#12134)
• [torch.compile] disable logging when cache is disabled ([torch.compile] disable logging when cache is disabled vllm-project/vllm#12043)
• [misc] fix cross-node TP ([misc] fix cross-node TP vllm-project/vllm#12166)
• [AMD][CI/Build][Bugfix] use pytorch stale wheel ([AMD][CI/Build][Bugfix] updated pytorch stale wheel path by using stable wheel vllm-project/vllm#12172)
• [core] further polish memory profiling ([core] further polish memory profiling vllm-project/vllm#12126)
• [Docs] Fix broken link in SECURITY.md ([Docs] Fix broken link in SECURITY.md vllm-project/vllm#12175)
• [Model] Port deepseek-vl2 processor, remove dependency ([Model] Port deepseek-vl2 processor and remove deepseek_vl2 dependency vllm-project/vllm#12169)
• [core] clean up executor class hierarchy between v1 and v0 ([core] clean up executor class hierarchy between v1 and v0 vllm-project/vllm#12171)
• [Misc] Support register quantization method out-of-tree ([Misc] Support register quantization method out-of-tree vllm-project/vllm#11969)
• [V1] Collect env var for usage stats ([V1] Collect env var for usage stats vllm-project/vllm#12115)
• [BUGFIX] Move scores to float32 in case of running xgrammar on cpu ([BUGFIX] Move scores to float32 in case of running xgrammar on cpu vllm-project/vllm#12152)
• [Bugfix] Fix multi-modal processors for transformers 4.48 ([Bugfix] Fix multi-modal processors for transformers 4.48 vllm-project/vllm#12187)
• [torch.compile] store inductor compiled Python file ([torch.compile] store inductor compiled Python file vllm-project/vllm#12182)
• benchmark_serving support --served-model-name param (benchmark_serving support --served-model-name param vllm-project/vllm#12109)
• [Misc] Add BNB support to GLM4-V model ([Misc] Add BNB support to GLM4-V model vllm-project/vllm#12184)
• [V1] Add V1 support of Qwen2-VL ([V1] Add V1 support of Qwen2-VL  vllm-project/vllm#12128)
• [Model] Support for fairseq2 Llama ([Model] Support for fairseq2 Llama vllm-project/vllm#11442)
• [Bugfix] Fix num_heads value for simple connector when tp enabled ([Bugfix] Fix num_heads value for simple connector when tp enabled vllm-project/vllm#12074)
• [torch.compile] fix sym_tensor_indices ([torch.compile] fix sym_tensor_indices vllm-project/vllm#12191)
• Move linting to pre-commit (Move linting to pre-commit vllm-project/vllm#11975)
• [DOC] Fix typo in docstring and assert message ([DOC] Fix typo in SingleStepOutputProcessor docstring and assert message vllm-project/vllm#12194)
• [DOC] Add missing docstring in LLMEngine.add_request() ([DOC] Add missing docstring for additional args in LLMEngine.add_request() vllm-project/vllm#12195)
• [Bugfix] Fix incorrect types in LayerwiseProfileResults ([Bugfix] Fix incorrect types in LayerwiseProfileResults vllm-project/vllm#12196)
• [Model] Add Qwen2 PRM model support ([Model] Add Qwen2 PRM model support vllm-project/vllm#12202)
• [Core] Interface for accessing model from VllmRunner ([Core] Interface for accessing model from VllmRunner vllm-project/vllm#10353)
• [misc] add placeholder format.sh ([misc] add placeholder format.sh vllm-project/vllm#12206)
• [CI/Build] Remove dummy CI steps ([CI/Build] Remove dummy CI steps vllm-project/vllm#12208)
• [CI/Build] Make pre-commit faster ([CI/Build] Make pre-commit faster vllm-project/vllm#12212)
• [Model] Upgrade Aria to transformers 4.48 ([Model] Upgrade Aria to transformers 4.48 vllm-project/vllm#12203)
• [misc] print a message to suggest how to bypass commit hooks ([misc] print a message to suggest how to bypass commit hooks vllm-project/vllm#12217)
• [core][bugfix] configure env var during import vllm ([core][bugfix] configure env var during import vllm vllm-project/vllm#12209)
• [V1] Remove _get_cache_block_size ([V1] Remove _get_cache_block_size vllm-project/vllm#12214)
• [Misc] Pass attention to impl backend ([Misc] Pass attention to impl backend vllm-project/vllm#12218)
• [Bugfix] Fix HfExampleModels.find_hf_info ([Bugfix] Fix HfExampleModels.find_hf_info vllm-project/vllm#12223)
• [CI] Pass local python version explicitly to pre-commit mypy.sh ([CI] Pass local python version explicitly to pre-commit mypy.sh vllm-project/vllm#12224)
• [Misc] Update CODEOWNERS ([Misc] Update CODEOWNERS vllm-project/vllm#12229)
• fix: update platform detection for M-series arm based MacBook processors (fix: update platform detection for M-series arm based MacBook processors vllm-project/vllm#12227)
• [misc] add cuda runtime version to usage data ([misc] add cuda runtime version to usage data vllm-project/vllm#12190)
• [bugfix] catch xgrammar unsupported array constraints ([bugfix] catch xgrammar unsupported array constraints vllm-project/vllm#12210)
• [Kernel] optimize moe_align_block_size for cuda graph and large num_experts (e.g. DeepSeek-V3) ([Kernel] optimize moe_align_block_size for cuda graph and large num_experts (e.g. DeepSeek-V3) vllm-project/vllm#12222)
• Add quantization and guided decoding CODEOWNERS (Add quantization and guided decoding CODEOWNERS vllm-project/vllm#12228)
• [AMD][Build] Porting dockerfiles from the ROCm/vllm fork ([AMD][Build] Porting dockerfiles from the ROCm/vllm fork vllm-project/vllm#11777)
• [BugFix] Fix GGUF tp>1 when vocab_size is not divisible by 64 ([BugFix] Fix GGUF tp>1 models when vocab_size is not divisible by 64 vllm-project/vllm#12230)
• [ci/build] disable failed and flaky tests ([ci/build] disable failed and flaky tests vllm-project/vllm#12240)
• [Misc] Rename MultiModalInputsV2 -> MultiModalInputs ([Misc] Rename MultiModalInputsV2 -> MultiModalInputs vllm-project/vllm#12244)
• [Misc]Add BNB quantization for PaliGemmaForConditionalGeneration ([Misc]Add BNB quantization for PaliGemmaForConditionalGeneration  vllm-project/vllm#12237)
• [Misc] Remove redundant TypeVar from base model ([Misc] Remove redundant TypeVar from base model vllm-project/vllm#12248)
• [Bugfix] Fix mm_limits access for merged multi-modal processor ([Bugfix] Fix mm_limits access for merged multi-modal processor vllm-project/vllm#12252)