[Core] LoRA V1 - Add add/pin/list/remove_lora functions (#13705)

This commit is contained in:
Varun Sundar Rabindranath
2025-02-25 00:18:02 -08:00
committed by GitHub
parent 4d251ad00e
commit 03f48b3db6
8 changed files with 270 additions and 22 deletions
+9 -4
View File
@@ -7,6 +7,7 @@ from typing import List
import pytest
from huggingface_hub import snapshot_download
import vllm.envs as env
from vllm.engine.arg_utils import AsyncEngineArgs
from vllm.inputs import TextPrompt
from vllm.lora.request import LoRARequest
@@ -144,10 +145,14 @@ async def test_add_lora():
await requests_processing_time(llm, dummy_run_requests)
# Run with warmup
for lr in warmup_run_requests:
await llm.add_lora(lr)
# Wait for the add_lora function to complete on the server side.
await asyncio.sleep(30)
add_lora_tasks = [llm.add_lora(lr) for lr in warmup_run_requests]
add_lora_results = await asyncio.gather(*add_lora_tasks)
if env.VLLM_USE_V1:
# Test that all all_lora calls are successful.
assert all(add_lora_results)
else:
# No way to check V0 engine results as the calls just return None.
pass
time_with_add_lora = await requests_processing_time(
llm, warmup_run_requests)