Remove hardcoded device="cuda" to support more devices (#2503)

Co-authored-by: Jiang Li <jiang1.li@intel.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
This commit is contained in:
Kunshang Ji
2024-02-01 15:46:39 -08:00
committed by GitHub
co-authored by Jiang Li
parent c410f5d020
commit 96b6f475dd
32 changed files with 343 additions and 292 deletions
+3 -1
View File
@@ -5,7 +5,8 @@ from unittest.mock import patch
from vllm.lora.models import LoRAMapping
from vllm.lora.request import LoRARequest
from vllm.config import ModelConfig, ParallelConfig, SchedulerConfig, LoRAConfig
from vllm.config import (ModelConfig, ParallelConfig, SchedulerConfig,
DeviceConfig, LoRAConfig)
from vllm.worker.worker import Worker
@@ -25,6 +26,7 @@ def test_worker_apply_lora(sql_lora_files):
),
parallel_config=ParallelConfig(1, 1, False),
scheduler_config=SchedulerConfig(32, 32, 32, 256),
device_config=DeviceConfig("cuda"),
local_rank=0,
rank=0,
lora_config=LoRAConfig(max_lora_rank=8, max_cpu_loras=32,