c721ae6042
### What this PR does / why we need it?
Use function CustomOp.register_oot to achieve the customop registery
```
from vllm.model_executor.custom_op import CustomOp
CustomOp.register_oot(_decorated_op_cls=AscendRMSNorm, name="RMSNorm")
```
### Does this PR introduce _any_ user-facing change?
N/A
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: v0.10.0
- vLLM main:
afa5b7ca0b
---------
Signed-off-by: Icey <1790571317@qq.com>
54 lines
1.9 KiB
Python
54 lines
1.9 KiB
Python
from unittest.mock import patch
|
|
|
|
import pytest
|
|
import torch
|
|
from vllm.model_executor.layers.layernorm import RMSNorm
|
|
|
|
|
|
@pytest.fixture
|
|
def dummy_tensor():
|
|
return torch.randn(4, 8, dtype=torch.float16)
|
|
|
|
|
|
def mock_rms_norm(x, weight, eps):
|
|
return x + 1, None
|
|
|
|
|
|
def mock_add_rms_norm(x, residual, weight, eps):
|
|
return 2 * x, None, 2 * residual
|
|
|
|
|
|
@pytest.mark.parametrize("is_310p_return", [True, False])
|
|
@pytest.mark.parametrize("residual",
|
|
[None, torch.randn(4, 8, dtype=torch.float32)])
|
|
@patch("torch_npu.npu_rms_norm", side_effect=mock_rms_norm)
|
|
@patch("torch_npu.npu_add_rms_norm", side_effect=mock_add_rms_norm)
|
|
def test_RMSNorm_forward(mock_add_rmsnorm, mock_rmsnorm, is_310p_return,
|
|
residual, dummy_tensor):
|
|
|
|
with patch("vllm_ascend.utils.is_310p", return_value=is_310p_return):
|
|
layer = RMSNorm(hidden_size=32, eps=1e-05)
|
|
if residual is not None:
|
|
out_x, out_residual = layer.forward_oot(dummy_tensor, residual)
|
|
|
|
if is_310p_return:
|
|
expected_arg_x = dummy_tensor + residual.to(dummy_tensor.dtype)
|
|
expected_out_x = expected_arg_x + 1
|
|
expected_out_residual = expected_arg_x.to(residual.dtype)
|
|
|
|
mock_rmsnorm.assert_called_once()
|
|
assert torch.allclose(out_x, expected_out_x)
|
|
assert torch.allclose(out_residual, expected_out_residual)
|
|
else:
|
|
expected_out_x = 2 * dummy_tensor
|
|
expected_out_residual = 2 * residual
|
|
mock_add_rmsnorm.assert_called_once()
|
|
assert torch.allclose(out_x, expected_out_x)
|
|
assert torch.allclose(out_residual, expected_out_residual)
|
|
else:
|
|
out_x = layer.forward(dummy_tensor, residual)
|
|
expected_out_x = dummy_tensor + 1
|
|
|
|
mock_rmsnorm.assert_called_once()
|
|
assert torch.allclose(out_x, expected_out_x)
|