[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>
This commit is contained in:
2026-10-19 21:06:19 +07:00
co-authored by Hinata
parent 5f3c7bd8ab
commit 5aa2ac64f8
269 changed files with 417 additions and 29932 deletions
+4 -124
View File
@@ -1,28 +1,16 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import json
import pytest
from tests.utils import VLLM_PATH
from vllm.engine.arg_utils import AsyncEngineArgs
from vllm.entrypoints.launchers.cli_args import (
make_arg_parser,
validate_parsed_serve_args,
)
from vllm.entrypoints.openai.models.protocol import LoRAModulePath
from vllm.exceptions import VLLMValidationError
from vllm.utils.argparse_utils import FlexibleArgumentParser
LORA_MODULE = {
"name": "module2",
"path": "/path/to/module2",
"base_model_name": "llama",
}
CHATML_JINJA_PATH = VLLM_PATH / "examples/template_chatml.jinja"
assert CHATML_JINJA_PATH.exists()
def _build_vllm_parsers():
vllm_parser = FlexibleArgumentParser()
@@ -96,87 +84,6 @@ def test_logging_config_cli_args(serve_parser):
assert config.pylogging_config_file == "/tmp/flat.json"
### Tests for LoRA module parsing
def test_valid_key_value_format(serve_parser):
# Test old format: name=path
args = serve_parser.parse_args(
[
"--lora-modules",
"module1=/path/to/module1",
]
)
expected = [LoRAModulePath(name="module1", path="/path/to/module1")]
assert args.lora_modules == expected
def test_valid_json_format(serve_parser):
# Test valid JSON format input
args = serve_parser.parse_args(
[
"--lora-modules",
json.dumps(LORA_MODULE),
]
)
expected = [
LoRAModulePath(name="module2", path="/path/to/module2", base_model_name="llama")
]
assert args.lora_modules == expected
def test_invalid_json_format(serve_parser):
# Test invalid JSON format input, missing closing brace
with pytest.raises(SystemExit):
serve_parser.parse_args(
["--lora-modules", '{"name": "module3", "path": "/path/to/module3"']
)
def test_invalid_type_error(serve_parser):
# Test type error when values are not JSON or key=value
with pytest.raises(SystemExit):
serve_parser.parse_args(
[
"--lora-modules",
"invalid_format", # This is not JSON or key=value format
]
)
def test_invalid_json_field(serve_parser):
# Test valid JSON format but missing required fields
with pytest.raises(SystemExit):
serve_parser.parse_args(
[
"--lora-modules",
'{"name": "module4"}', # Missing required 'path' field
]
)
def test_empty_values(serve_parser):
# Test when no LoRA modules are provided
args = serve_parser.parse_args(["--lora-modules", ""])
assert args.lora_modules == []
def test_multiple_valid_inputs(serve_parser):
# Test multiple valid inputs (both old and JSON format)
args = serve_parser.parse_args(
[
"--lora-modules",
"module1=/path/to/module1",
json.dumps(LORA_MODULE),
]
)
expected = [
LoRAModulePath(name="module1", path="/path/to/module1"),
LoRAModulePath(
name="module2", path="/path/to/module2", base_model_name="llama"
),
]
assert args.lora_modules == expected
### Tests for serve argument validation that run prior to loading
@@ -241,10 +148,12 @@ def test_reasoning_parser_none_disables_parser(serve_parser, value, expected):
assert args.reasoning_parser == expected
def test_chat_template_validation_for_happy_paths(serve_parser):
def test_chat_template_validation_for_happy_paths(serve_parser, tmp_path):
"""Ensure validation passes if the chat template exists."""
chat_template = tmp_path / "template.jinja"
chat_template.write_text("{{ messages }}")
args = serve_parser.parse_args(
args=["--chat-template", CHATML_JINJA_PATH.absolute().as_posix()]
args=["--chat-template", chat_template.absolute().as_posix()]
)
validate_parsed_serve_args(args)
@@ -389,32 +298,3 @@ def test_served_model_name_parsing(tmp_path, vllm_parser, args, raises):
else:
with pytest.raises(raises):
vllm_parser.parse_args(args=args)
### Tests for LoRA target modules parsing
def test_lora_target_modules_single(serve_parser):
"""Test parsing single lora-target-modules argument."""
args = serve_parser.parse_args(
args=["--enable-lora", "--lora-target-modules", "o_proj"]
)
assert args.lora_target_modules == ["o_proj"]
def test_lora_target_modules_multiple(serve_parser):
"""Test parsing multiple lora-target-modules arguments."""
args = serve_parser.parse_args(
args=[
"--enable-lora",
"--lora-target-modules",
"o_proj",
"qkv_proj",
"down_proj",
]
)
assert args.lora_target_modules == ["o_proj", "qkv_proj", "down_proj"]
def test_lora_target_modules_default_none(serve_parser):
"""Test that lora-target-modules defaults to None."""
args = serve_parser.parse_args(args=[])
assert args.lora_target_modules is None