64 lines
2.0 KiB
Python
64 lines
2.0 KiB
Python
"""
|
|
Copyright 2023-2024 SGLang Team
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
"""
|
|
|
|
from typing import Optional, Tuple, Union
|
|
|
|
import torch
|
|
import torch.nn as nn
|
|
from flashinfer.norm import fused_add_rmsnorm, rmsnorm
|
|
from vllm.model_executor.custom_op import CustomOp
|
|
|
|
|
|
class RMSNorm(CustomOp):
|
|
def __init__(
|
|
self,
|
|
hidden_size: int,
|
|
eps: float = 1e-6,
|
|
) -> None:
|
|
super().__init__()
|
|
self.weight = nn.Parameter(torch.ones(hidden_size))
|
|
self.variance_epsilon = eps
|
|
|
|
def forward_cuda(
|
|
self,
|
|
x: torch.Tensor,
|
|
residual: Optional[torch.Tensor] = None,
|
|
) -> Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]:
|
|
|
|
if residual is not None:
|
|
fused_add_rmsnorm(x, residual, self.weight.data, self.variance_epsilon)
|
|
return x, residual
|
|
out = rmsnorm(x, self.weight.data, self.variance_epsilon)
|
|
return out
|
|
|
|
def forward_native(
|
|
self,
|
|
x: torch.Tensor,
|
|
residual: Optional[torch.Tensor] = None,
|
|
) -> Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]:
|
|
orig_dtype = x.dtype
|
|
x = x.to(torch.float32)
|
|
if residual is not None:
|
|
x = x + residual.to(torch.float32)
|
|
residual = x.to(orig_dtype)
|
|
|
|
variance = x.pow(2).mean(dim=-1, keepdim=True)
|
|
x = x * torch.rsqrt(variance + self.variance_epsilon)
|
|
x = x.to(orig_dtype) * self.weight
|
|
if residual is None:
|
|
return x
|
|
else:
|
|
return x, residual
|