Skip to content

Improve error handling for requests with unloaded LoRA path(s) #7642

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Jul 2, 2025
Merged

Improve error handling for requests with unloaded LoRA path(s) #7642

Show file tree
Hide file tree
Changes from 6 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
9f22632
Handle requests with unloaded LoRA.
lifuhuang Jun 30, 2025
3899b29
Minor
lifuhuang Jun 30, 2025
8139e8d
Merge remote-tracking branch 'origin/main' into lifuhuang/lora-error
lifuhuang Jun 30, 2025
2db1b9a
Fix typo.
lifuhuang Jun 30, 2025
dae680d
Fix
lifuhuang Jun 30, 2025
e838b47
Merge branch 'main' into lifuhuang/lora-error
lifuhuang Jun 30, 2025
3dc0509
Merge branch 'main' into lifuhuang/lora-error
lifuhuang Jun 30, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 26 additions & 0 deletions python/sglang/srt/managers/tokenizer_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -242,6 +242,12 @@ def __init__(
revision=server_args.revision,
)

# Initialize loaded loRA adapters with the initial lora paths in the server_args.
# This list will be updated when new LoRA adapters are loaded or unloaded dynamically.
self.loaded_lora_adapters: Dict[str, str] = dict(
self.server_args.lora_paths or {}
)

# Store states
self.no_create_loop = False
self.rid_to_state: Dict[str, ReqState] = {}
Expand Down Expand Up @@ -547,6 +553,9 @@ def _validate_one_request(
"Please set `--enable-custom-logits-processor` to enable this feature."
)

if self.server_args.lora_paths and obj.lora_path:
self._validate_lora_adapters(obj)

def _create_tokenized_object(
self,
obj: Union[GenerateReqInput, EmbeddingReqInput],
Expand Down Expand Up @@ -651,6 +660,21 @@ def _validate_batch_tokenization_constraints(
"Batch tokenization is not needed for input_embeds. Do not set `enable_tokenizer_batch_encode`."
)

def _validate_lora_adapters(self, obj: GenerateReqInput):
"""Validate that the requested LoRA adapters are loaded."""
requested_adapters = (
set(obj.lora_path) if isinstance(obj.lora_path, list) else {obj.lora_path}
)
loaded_adapters = (
self.loaded_lora_adapters.keys() if self.loaded_lora_adapters else set()
)
unloaded_adapters = requested_adapters - loaded_adapters
if unloaded_adapters:
raise ValueError(
f"The following requested LoRA adapters are not loaded: {unloaded_adapters}\n"
f"Loaded adapters: {loaded_adapters}."
)

def _send_one_request(
self,
obj: Union[GenerateReqInput, EmbeddingReqInput],
Expand Down Expand Up @@ -977,6 +1001,7 @@ async def load_lora_adapter(

async with self.model_update_lock.writer_lock:
result = (await self.update_lora_adapter_communicator(obj))[0]
self.loaded_lora_adapters = result.loaded_adapters
return result

async def unload_lora_adapter(
Expand All @@ -998,6 +1023,7 @@ async def unload_lora_adapter(

async with self.model_update_lock.writer_lock:
result = (await self.update_lora_adapter_communicator(obj))[0]
self.loaded_lora_adapters = result.loaded_adapters
return result

async def get_weights_by_name(
Expand Down
4 changes: 2 additions & 2 deletions python/sglang/srt/server_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@
import os
import random
import tempfile
from typing import List, Literal, Optional
from typing import List, Literal, Optional, Union

from sglang.srt.hf_transformers_utils import check_gguf_file, get_config
from sglang.srt.reasoning_parser import ReasoningParser
Expand Down Expand Up @@ -130,7 +130,7 @@ class ServerArgs:
preferred_sampling_params: Optional[str] = None

# LoRA
lora_paths: Optional[List[str]] = None
lora_paths: Optional[Union[dict[str, str], List[str]]] = None
max_loras_per_batch: int = 8
lora_backend: str = "triton"

Expand Down
126 changes: 108 additions & 18 deletions test/srt/models/lora/test_lora_update.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@
import unittest
from dataclasses import dataclass
from enum import Enum
from typing import List, Optional, Union
from typing import Any, List, Optional, Union

import requests
import torch
Expand All @@ -42,14 +42,16 @@
class OperationType(Enum):
LOAD = "load"
UNLOAD = "unload"
NOOP = "noop"
FORWARD = "forward"
EXPECT_ERROR = "expect_error"


@dataclass
class Operation:
# Operation type, can be LOAD, UNLOAD, FORWARD, or EXPECT_ERROR
type: OperationType
data: Optional[str]
# Data associated with the operation. Exact type varies depending on the operation
data: Optional[Any]


@dataclass
Expand All @@ -62,7 +64,7 @@ class TestCase:
max_new_tokens: int = 32


def create_batch_data(adapters: Union[str, list]) -> dict:
def create_batch_data(adapters: Union[str, list]) -> List[tuple[str, str]]:
if not isinstance(adapters, list):
adapters = [adapters]
return [(prompt, adapter) for prompt in PROMPTS for adapter in adapters]
Expand All @@ -80,6 +82,26 @@ def create_batch_data(adapters: Union[str, list]) -> dict:
],
initial_adapters=["philschmid/code-llama-3-1-8b-text-to-sql-lora"],
op_sequence=[
Operation(
type=OperationType.FORWARD,
data=create_batch_data("philschmid/code-llama-3-1-8b-text-to-sql-lora"),
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data(
"Nutanix/Meta-Llama-3.1-8B-Instruct_lora_4_alpha_16"
),
"not loaded",
),
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data("pbevan11/llama-3.1-8b-ocr-correction"),
"not loaded",
),
),
Operation(
type=OperationType.LOAD,
data="Nutanix/Meta-Llama-3.1-8B-Instruct_lora_4_alpha_16",
Expand All @@ -102,6 +124,13 @@ def create_batch_data(adapters: Union[str, list]) -> dict:
type=OperationType.UNLOAD,
data="philschmid/code-llama-3-1-8b-text-to-sql-lora",
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data("philschmid/code-llama-3-1-8b-text-to-sql-lora"),
"not loaded",
),
),
Operation(
type=OperationType.FORWARD,
data=create_batch_data(
Expand All @@ -115,6 +144,15 @@ def create_batch_data(adapters: Union[str, list]) -> dict:
type=OperationType.UNLOAD,
data="Nutanix/Meta-Llama-3.1-8B-Instruct_lora_4_alpha_16",
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data(
"Nutanix/Meta-Llama-3.1-8B-Instruct_lora_4_alpha_16"
),
"not loaded",
),
),
Operation(
type=OperationType.FORWARD,
data=create_batch_data("pbevan11/llama-3.1-8b-ocr-correction"),
Expand Down Expand Up @@ -149,6 +187,22 @@ def create_batch_data(adapters: Union[str, list]) -> dict:
type=OperationType.FORWARD,
data=create_batch_data("philschmid/code-llama-3-1-8b-text-to-sql-lora"),
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data(
"Nutanix/Meta-Llama-3.1-8B-Instruct_lora_4_alpha_16"
),
"not loaded",
),
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data("pbevan11/llama-3.1-8b-ocr-correction"),
"not loaded",
),
),
Operation(
type=OperationType.LOAD,
data="pbevan11/llama-3.1-8b-ocr-correction",
Expand All @@ -157,6 +211,13 @@ def create_batch_data(adapters: Union[str, list]) -> dict:
type=OperationType.UNLOAD,
data="philschmid/code-llama-3-1-8b-text-to-sql-lora",
),
Operation(
type=OperationType.EXPECT_ERROR,
data=(
create_batch_data("philschmid/code-llama-3-1-8b-text-to-sql-lora"),
"not loaded",
),
),
Operation(
type=OperationType.FORWARD,
data=create_batch_data("pbevan11/llama-3.1-8b-ocr-correction"),
Expand Down Expand Up @@ -332,19 +393,31 @@ def forward(
prompts: List[str],
lora_paths: List[str],
max_new_tokens: int = 32,
expected_error: str = None,
):
"""
Perform a batch forward pass with the current set of loaded LoRA adapters.
"""
response = self.handle.batch_forward(
prompts=prompts,
lora_paths=lora_paths,
max_new_tokens=max_new_tokens,
)
output_strs = response.output_strs
try:
response = self.handle.batch_forward(
prompts=prompts,
lora_paths=lora_paths,
max_new_tokens=max_new_tokens,
)
except ValueError as e:
if expected_error:
error_message = str(e)
self.testcase.assertIn(expected_error, error_message)
print(f"Received error as expected: {error_message}")
return error_message

raise e

self.testcase.assertEqual(len(response.output_strs), len(prompts))
output = response.output_strs
print(f"output_strs: {output}")

print(f"output_strs: {output_strs}")
return output_strs
return output


class LoRAUpdateServerTestSession(LoRAUpdateTestSessionBase):
Expand Down Expand Up @@ -426,6 +499,7 @@ def forward(
prompts: List[str],
lora_paths: List[str],
max_new_tokens: int = 32,
expected_error: str = None,
):
"""
Perform a batch forward pass with the current set of loaded LoRA adapters.
Expand All @@ -442,11 +516,18 @@ def forward(
},
},
)
self.testcase.assertTrue(response.ok)
output_strs = [r["text"] for r in response.json()]

print(f"output_strs: {output_strs}")
return output_strs
if expected_error:
self.testcase.assertEqual(response.status_code, 400)
self.testcase.assertIn(expected_error, response.text)
output = response.text
print(f"Received error as expected: {response.text}")
return output
else:
self.testcase.assertTrue(response.ok)
output = [r["text"] for r in response.json()]
self.testcase.assertEqual(len(output), len(prompts))
print(f"output_strs: {output}")
return output


# Factory function to create the appropriate LoRA test session based on mode
Expand Down Expand Up @@ -535,14 +616,23 @@ def _run_operation_sequence(
max_new_tokens=max_new_tokens,
)
forward_outputs.append(result)
elif op_type == OperationType.EXPECT_ERROR:
input_data, expected_error = data
prompts, adapters = zip(*input_data)
result = session.forward(
prompts=list(prompts),
lora_paths=list(adapters),
max_new_tokens=max_new_tokens,
expected_error=expected_error,
)

return forward_outputs

def test_dynamic_adapter_updates(self):
for case_idx, test_case in enumerate(TEST_CASES, start=1):
for mode in [
LoRAUpdateTestSessionMode.SERVER,
LoRAUpdateTestSessionMode.ENGINE,
LoRAUpdateTestSessionMode.SERVER,
]:
print("=" * 100)
print(f"Starting test case {case_idx} in {mode.value} mode.")
Expand Down
Loading