[Chore] Clean up pytorch helper functions in vllm.utils (#26908)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn> Signed-off-by: isotr0py <2037008807@qq.com>
This commit is contained in:
@@ -3,7 +3,10 @@
|
||||
|
||||
import pytest
|
||||
|
||||
from vllm.utils import create_kv_caches_with_random, create_kv_caches_with_random_flash
|
||||
from vllm.utils.torch_utils import (
|
||||
create_kv_caches_with_random,
|
||||
create_kv_caches_with_random_flash,
|
||||
)
|
||||
|
||||
|
||||
@pytest.fixture()
|
||||
|
||||
@@ -15,7 +15,7 @@ from tests.kernels.utils import make_alibi_bias
|
||||
from vllm.attention.ops.chunked_prefill_paged_decode import chunked_prefill_paged_decode
|
||||
from vllm.attention.ops.prefix_prefill import context_attention_fwd
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import STR_DTYPE_TO_TORCH_DTYPE
|
||||
from vllm.utils.torch_utils import STR_DTYPE_TO_TORCH_DTYPE
|
||||
|
||||
NUM_HEADS = [64]
|
||||
NUM_QUERIES_PER_KV = [1, 64]
|
||||
|
||||
@@ -3,7 +3,8 @@
|
||||
import pytest
|
||||
import torch
|
||||
|
||||
from vllm.utils import get_cuda_view_from_cpu_tensor, is_uva_available
|
||||
from vllm.utils import is_uva_available
|
||||
from vllm.utils.torch_utils import get_cuda_view_from_cpu_tensor
|
||||
|
||||
CUDA_DEVICES = [f"cuda:{i}" for i in range(1 if torch.cuda.device_count() == 1 else 2)]
|
||||
|
||||
|
||||
@@ -13,8 +13,9 @@ import torch
|
||||
import vllm.model_executor.layers.fused_moe.modular_kernel as mk
|
||||
from vllm.config import VllmConfig, set_current_vllm_config
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import cuda_device_count_stateless, has_deep_ep, has_deep_gemm, has_pplx
|
||||
from vllm.utils import has_deep_ep, has_deep_gemm, has_pplx
|
||||
from vllm.utils.flashinfer import has_flashinfer_cutlass_fused_moe
|
||||
from vllm.utils.torch_utils import cuda_device_count_stateless
|
||||
|
||||
from .modular_kernel_tools.common import (
|
||||
Config,
|
||||
|
||||
@@ -22,8 +22,8 @@ from vllm.utils import (
|
||||
STR_BACKEND_ENV_VAR,
|
||||
STR_FLASH_ATTN_VAL,
|
||||
STR_XFORMERS_ATTN_VAL,
|
||||
make_tensor_with_pad,
|
||||
)
|
||||
from vllm.utils.torch_utils import make_tensor_with_pad
|
||||
|
||||
# For now, disable "test_aot_dispatch_dynamic" since there are some
|
||||
# bugs related to this test in PyTorch 2.4.
|
||||
|
||||
Reference in New Issue
Block a user