Files
MOVZXandHinata 5aa2ac64f8 [Mini-VLLM] Remove LoRA, plugins, speech-to-text and scale-out surfaces
Production runs one static model behind the OpenAI-compatible API. Every
adapter path and non-OpenAI HTTP surface is unreachable there.

- Delete the LoRA stack: vllm/lora, LoRARequest on the engine wire, the
  scheduler, worker, cudagraph and MoE plumbing, LoRA metrics, the lora
  CLI group, and the LoRA tests.
- Drop the plugin loader in vllm/plugins. Platform resolution is now
  CUDA-only and the general/stat-logger plugin hooks are gone.
- Delete the speech-to-text and realtime entrypoints, the scale-out
  render routes, and vllm/config/speech_to_text.py. GenerationTask is
  generate-only.
- Cut DecodeBenchConnector and the disagg-prefill doc and image. The
  Multi, Offloading and Example connectors stay; production uses the
  CPU KV offload tier.
- Delete the ROCm-only fused KDA chunk kernel (2346 lines) with its CMake
  arch gate, declarations and op registrations. The CUDA build never
  defined the guard and no Python caller existed.
- Delete the orphan Mistral and Pixtral chat fixtures and examples/.
- Fix _check_model: removing the LoRA early return made every request
  fail with "The model ... does not exist. Valid aliases: ...".

Verified on an RTX 3090: import sweep reports 1242 modules with 7 known
optional-dependency failures; a live server answers chat, completions,
responses, tokenize and detokenize with zero log errors and 404 for the
removed surfaces; collection yields 8558 tests with the 3 pre-existing
errors; ruff adds no new finding; the incremental _C build succeeds.

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-19 21:06:19 +07:00

152 lines
4.5 KiB
Python

# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from vllm.config.attention import AttentionConfig, HiSparseConfig
from vllm.config.aux_output import AuxOutputConfig
from vllm.config.cache import CacheConfig
from vllm.config.compilation import (
CompilationConfig,
CompilationMode,
CUDAGraphMode,
PassConfig,
)
from vllm.config.device import DeviceConfig
from vllm.config.diffusion import DiffusionConfig
from vllm.config.ec_manager_config import EncoderCacheManagerConfig
from vllm.config.ec_transfer import ECTransferConfig
from vllm.config.engram import EngramConfig
from vllm.config.fault_tolerance import FaultToleranceConfig
from vllm.config.kernel import KernelConfig
from vllm.config.kv_events import KVEventsConfig
from vllm.config.kv_transfer import KVTransferConfig
from vllm.config.load import LoadConfig
from vllm.config.logging import LoggingConfig
from vllm.config.mamba import MambaConfig
from vllm.config.model import (
ModelConfig,
iter_architecture_defaults,
str_dtype_to_torch_dtype,
try_match_architecture_defaults,
)
from vllm.config.multimodal import MultiModalConfig
from vllm.config.observability import ObservabilityConfig
from vllm.config.offload import (
OffloadBackend,
OffloadConfig,
PrefetchOffloadConfig,
UVAOffloadConfig,
)
from vllm.config.parallel import EPLBConfig, ParallelConfig
from vllm.config.pooler import PoolerConfig
from vllm.config.profiler import ProfilerConfig
from vllm.config.reasoning import ReasoningConfig
from vllm.config.scheduler import SchedulerConfig
from vllm.config.speculative import SpeculativeConfig
from vllm.config.structured_outputs import StructuredOutputsConfig
from vllm.config.utils import (
ConfigType,
SupportsMetricsInfo,
config,
get_attr_docs,
is_init_field,
replace,
update_config,
)
from vllm.config.vllm import (
VllmConfig,
get_cached_compilation_config,
get_current_vllm_config,
get_current_vllm_config_or_none,
get_layers_from_vllm_config,
set_current_vllm_config,
)
from vllm.config.watermarking import WatermarkConfig
from vllm.config.weight_transfer import WeightTransferConfig
# __all__ should only contain classes and functions.
# Types and globals should be imported from their respective modules.
__all__ = [
# From vllm.config.aux_output
"AuxOutputConfig",
# From vllm.config.attention
"AttentionConfig",
"HiSparseConfig",
# From vllm.config.cache
"CacheConfig",
# From vllm.config.compilation
"CompilationConfig",
"CompilationMode",
"CUDAGraphMode",
"PassConfig",
# From vllm.config.device
"DeviceConfig",
# From vllm.config.diffusion
"DiffusionConfig",
# From vllm.config.ec_manager_config
"EncoderCacheManagerConfig",
# From vllm.config.ec_transfer
"ECTransferConfig",
# From vllm.config.engram
"EngramConfig",
# From vllm.config.kernel
"KernelConfig",
# From vllm.config.kv_events
"KVEventsConfig",
# From vllm.config.kv_transfer
"KVTransferConfig",
# From vllm.config.load
"LoadConfig",
# From vllm.config.logging
"LoggingConfig",
# From vllm.config.mamba
"MambaConfig",
# From vllm.config.model
"ModelConfig",
"iter_architecture_defaults",
"str_dtype_to_torch_dtype",
"try_match_architecture_defaults",
# From vllm.config.multimodal
"MultiModalConfig",
# From vllm.config.observability
"ObservabilityConfig",
# From vllm.config.offload
"OffloadBackend",
"OffloadConfig",
"PrefetchOffloadConfig",
"UVAOffloadConfig",
# From vllm.config.parallel
"EPLBConfig",
"ParallelConfig",
# From vllm.config.pooler
"PoolerConfig",
# From vllm.config.reasoning
"ReasoningConfig",
# From vllm.config.scheduler
"SchedulerConfig",
# From vllm.config.speculative
"SpeculativeConfig",
# From vllm.config.structured_outputs
"StructuredOutputsConfig",
# From vllm.config.profiler
"ProfilerConfig",
# From vllm.config.fault_tolerance
"FaultToleranceConfig",
# From vllm.config.utils
"ConfigType",
"SupportsMetricsInfo",
"config",
"get_attr_docs",
"is_init_field",
"replace",
"update_config",
# From vllm.config.vllm
"VllmConfig",
"get_cached_compilation_config",
"get_current_vllm_config",
"get_current_vllm_config_or_none",
"set_current_vllm_config",
"get_layers_from_vllm_config",
"WeightTransferConfig",
"WatermarkConfig",
]