EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

[CustomOp] Register RotaryEmbedding instead of overwrite forward (#2385)

Browse Source 
### What this PR does / why we need it?
Register RotaryEmbedding instead of overwrite forward

### 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:
808d2e9aa0

---------

Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
This commit is contained in:
Icey
2025-08-25 09:32:35 +08:00
committed by GitHub
parent 950c4b219a
commit f796e6280b
6 changed files with 426 additions and 381 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
vllm_ascend/ops/__init__.py
View File

@@ -17,12 +17,13 @@
import torch
import vllm_ascend.ops.activation # noqa
import vllm_ascend.ops.common_fused_moe # noqa
import vllm_ascend.ops.fused_moe # noqa
import vllm_ascend.ops.layernorm # noqa
import vllm_ascend.ops.rotary_embedding # noqa
import vllm_ascend.ops.vocab_parallel_embedding # noqa
from vllm_ascend.ops.activation import AscendQuickGELU, AscendSiluAndMul
from vllm_ascend.ops.rotary_embedding import (
AscendDeepseekScalingRotaryEmbedding, AscendRotaryEmbedding)
class dummyFusionOp:
@@ -47,3 +48,9 @@ def register_dummy_fusion_op() -> None:
name="fused_add_rms_norm_static_fp8_quant")
torch.ops._C.rms_norm_dynamic_per_token_quant = dummyFusionOp(
name="rms_norm_dynamic_per_token_quant")
__all__ = [
"AscendQuickGELU", "AscendSiluAndMul", "AscendRotaryEmbedding",
"AscendDeepseekScalingRotaryEmbedding"
]

539
vllm_ascend/ops/rotary_embedding.py
View File

@@ -25,6 +25,7 @@ from vllm.model_executor.layers.rotary_embedding import (
DeepseekScalingRotaryEmbedding, RotaryEmbedding)
from vllm_ascend.ascend_config import get_ascend_config
from vllm_ascend.platform import NPUPlatform
from vllm_ascend.utils import enable_custom_op, is_310p
@@ -89,167 +90,7 @@ def rope_forward_oot(
return query.view(query_shape), key.view(key_shape)
def native_rope_deepseek_forward(self,
positions: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
offsets: Optional[torch.Tensor] = None,
max_seq_len: Optional[int] = None):
if max_seq_len is not None and max_seq_len > self.max_seq_len:
_set_cos_sin_cache(self, max_seq_len, query.device, query.dtype)
if len(key.shape) == 2:
key = key[:, None, :]
# Note: we implement the non neox_style method with shuffle the last dim and neox style
# calculation method which is also more compute friendly to the ascend machine
# https://huggingface.co/deepseek-ai/DeepSeek-V3-0324/blob/main/modeling_deepseek.py
neox_style = True
if self.is_neox_style is False:
b, h_q, d = query.shape
query = query.view(b, h_q, d // 2, 2).transpose(3,
2).reshape(b, h_q, d)
b, h_k, d = key.shape
key = key.view(b, h_k, d // 2, 2).transpose(3, 2).reshape(b, h_k, d)
q_pe, k_pe = rope_forward_oot(self, positions, query, key, offsets,
neox_style)
return q_pe, k_pe
def rotate_half(x):
"""Rotates half the hidden dims of the input."""
x1 = x[..., :x.shape[-1] // 2]
x2 = x[..., x.shape[-1] // 2:]
return torch.cat((-x2, x1), dim=-1)
# Inverse dim formula to find dim based on number of rotations
def yarn_find_correction_dim(num_rotations,
dim,
base=10000,
max_position_embeddings=2048):
# Note: use torch instead of math to solve MTP compilation error.
return (dim * torch.log(
torch.tensor(max_position_embeddings) /
(num_rotations * 2 * torch.pi))) / (2 * torch.log(torch.tensor(base)))
def yarn_get_mscale(scale: float = 1, mscale: float = 1) -> float:
if scale <= 1:
return 1.0
return 0.1 * mscale * math.log(scale) + 1.0
# Find dim range bounds based on rotations
def yarn_find_correction_range(low_rot,
high_rot,
dim,
base=10000,
max_position_embeddings=2048):
# Note: use torch instead of math to solve MTP compilation error.
low = torch.floor(
yarn_find_correction_dim(low_rot, dim, base, max_position_embeddings))
high = torch.ceil(
yarn_find_correction_dim(high_rot, dim, base, max_position_embeddings))
# Note: use torch instead of max/min to solve MTP compilation error.
return torch.clamp(low, min=0), torch.clamp(high, max=dim - 1)
def yarn_linear_ramp_mask(min_value, max_value, dim):
# Note: The if conditional branch is not used here
# to solve MTP compilation error.
max_value += (min_value == max_value).float() * 0.001
linear_func = (torch.arange(dim, dtype=torch.float32) -
min_value) / (max_value - min_value)
ramp_func = torch.clamp(linear_func, 0, 1)
return ramp_func
# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
def apply_rotary_pos_emb(q, k, cos, sin, position_ids, unsqueeze_dim=1):
"""Applies Rotary Position Embedding to the query and key tensors.
Args:
q (`torch.Tensor`): The query tensor.
k (`torch.Tensor`): The key tensor.
cos (`torch.Tensor`): The cosine part of the rotary embedding.
sin (`torch.Tensor`): The sine part of the rotary embedding.
position_ids (`torch.Tensor`):
The position indices of the tokens corresponding to the query and key tensors. For example, this can be
used to pass offsetted position ids when working with a KV-cache.
unsqueeze_dim (`int`, *optional*, defaults to 1):
The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
Returns:
`tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
"""
cos = cos[position_ids]
sin = sin[position_ids]
cos = cos[:, None, None, :]
sin = sin[:, None, None, :]
if len(q.shape) == 3:
q = q[:, :, None, :]
if len(k.shape) == 2:
k = k[:, None, None, :]
elif len(k.shape) == 3:
k = k[:, :, None, :]
b, h_q, s, d = q.shape
q = q.view(b, h_q, s, d // 2, 2).transpose(4, 3).reshape(b, h_q, s, d)
b, h_k, s, d = k.shape
k = k.view(b, h_k, s, d // 2, 2).transpose(4, 3).reshape(b, h_k, s, d)
q_embed = (q * cos) + (rotate_half(q) * sin)
k_embed = (k * cos) + (rotate_half(k) * sin)
q_embed = q_embed.view(b, h_q, d)
k_embed = k_embed.view(b, h_k, d)
return q_embed, k_embed
def _set_cos_sin_cache(self, seq_len, device, dtype):
self.max_seq_len_cached = seq_len
dim = self.rotary_dim
freq_extra = 1.0 / (self.base**(
torch.arange(0, dim, 2, dtype=torch.float32, device=device) / dim))
freq_inter = 1.0 / (self.scaling_factor * self.base**(
torch.arange(0, dim, 2, dtype=torch.float32, device=device) / dim))
low, high = yarn_find_correction_range(
self.beta_fast,
self.beta_slow,
dim,
self.base,
self.max_position_embeddings,
)
inv_freq_mask = 1.0 - yarn_linear_ramp_mask(low, high, dim // 2).to(
device=device, dtype=torch.float32)
inv_freq = freq_inter * (1 - inv_freq_mask) + freq_extra * inv_freq_mask
self.register_buffer("inv_freq", inv_freq, persistent=False)
t = torch.arange(seq_len * self.scaling_factor,
device=device,
dtype=torch.float32)
freqs = torch.outer(t, inv_freq)
cos_cached = torch.cat([freqs, freqs], dim=-1).cos() * self.mscale
sin_cached = torch.cat([freqs, freqs], dim=-1).sin() * self.mscale
cos_cached = cos_cached.to(dtype)
sin_cached = sin_cached.to(dtype)
cache = torch.cat([freqs.cos() * self.mscale,
freqs.sin() * self.mscale],
dim=-1).to(dtype)
self.register_buffer("cos_sin_cache", cache, persistent=False)
self.register_buffer("cos_cached", cos_cached, persistent=False)
self.register_buffer("sin_cached", sin_cached, persistent=False)
def __set_cos_sin_cache(self, seq_len, device, dtype):
def set_cos_sin_cache(self, seq_len, device, dtype):
inv_freq = 1.0 / (self.base**(torch.arange(
0, self.rotary_dim, 2, device=device, dtype=torch.float32) *
(1 / self.rotary_dim)))
@@ -266,117 +107,275 @@ def __set_cos_sin_cache(self, seq_len, device, dtype):
self.embed = F.embedding
_original_re_init = RotaryEmbedding.__init__
class AscendRotaryEmbedding(RotaryEmbedding):
def qwen_rope_init_func(
self,
head_size: int,
rotary_dim: int,
max_position_embeddings: int,
base: float,
is_neox_style: bool,
dtype: torch.dtype,
) -> None:
_original_re_init(self, head_size, rotary_dim, max_position_embeddings,
base, is_neox_style, dtype)
if get_ascend_config().torchair_graph_config.enabled:
__set_cos_sin_cache(self,
seq_len=max_position_embeddings,
device="npu",
dtype=dtype)
def rope_forward(
self,
positions: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
offsets: Optional[torch.Tensor] = None,
is_neox_style_override: Optional[bool] = None,
max_seq_len: Optional[int] = None,
is_prefill: Optional[bool] = True,
is_qwen_torchair: Optional[bool] = False,
):
if get_ascend_config().torchair_graph_config.enabled \
and is_qwen_torchair and not is_prefill:
if max_seq_len is not None and torch.gt(max_seq_len,
self.max_position_embeddings):
__set_cos_sin_cache(self,
seq_len=max_seq_len,
device=query.device,
dtype=torch.float32)
# bsnd/bnsd
if positions is not None:
cos = self.embed(positions, self.cos)
sin = self.embed(positions, self.sin)
self.cos_embed = cos
self.sin_embed = sin
else:
cos = self.cos_embed
sin = self.sin_embed
query = query.view(*query.shape[:-1], -1, self.head_size).contiguous()
key = key.view(*key.shape[:-1], -1, self.head_size).contiguous()
cos = cos.unsqueeze(-2).unsqueeze(-2)
sin = sin.unsqueeze(-2).unsqueeze(-2)
query = query.unsqueeze(1)
key = key.unsqueeze(1)
q_embed, k_embed = torch_npu.npu_apply_rotary_pos_emb(
query, key, cos, sin)
return q_embed.flatten(-2), k_embed.flatten(-2)
else:
return rope_forward_oot(self, positions, query, key, offsets,
is_neox_style_override,
is_qwen_torchair) # type: ignore
def deepseek_rope_init_func(
self,
head_size: int,
rotary_dim: int,
max_position_embeddings: int,
base: int,
is_neox_style: bool,
scaling_factor: float,
dtype: torch.dtype,
*,
extrapolation_factor: float = 1,
attn_factor: float = 1,
beta_fast: int = 32,
beta_slow: int = 1,
mscale: float = 1,
mscale_all_dim: float = 0,
) -> None:
self.scaling_factor = scaling_factor
self.extrapolation_factor = extrapolation_factor
self.attn_factor = attn_factor
self.beta_fast = beta_fast
self.beta_slow = beta_slow
# Get n-d magnitude scaling corrected for interpolation.
self.mscale = float(
yarn_get_mscale(self.scaling_factor, float(mscale)) /
yarn_get_mscale(self.scaling_factor, float(mscale_all_dim)) *
attn_factor)
super(DeepseekScalingRotaryEmbedding,
self).__init__(head_size, rotary_dim, max_position_embeddings, base,
def __init__(
self,
head_size: int,
rotary_dim: int,
max_position_embeddings: int,
base: float,
is_neox_style: bool,
dtype: torch.dtype,
) -> None:
super().__init__(head_size, rotary_dim, max_position_embeddings, base,
is_neox_style, dtype)
self.max_seq_len = max_position_embeddings
_set_cos_sin_cache(self,
max_position_embeddings,
dtype=dtype,
device="npu")
if get_ascend_config().torchair_graph_config.enabled:
set_cos_sin_cache(self,
seq_len=max_position_embeddings,
device="npu",
dtype=dtype)
def forward_oot(
self,
positions: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
offsets: Optional[torch.Tensor] = None,
is_neox_style_override: Optional[bool] = None,
max_seq_len: Optional[int] = None,
is_prefill: Optional[bool] = True,
is_qwen_torchair: Optional[bool] = False,
):
if get_ascend_config().torchair_graph_config.enabled \
and is_qwen_torchair and not is_prefill:
if max_seq_len is not None and torch.gt(
max_seq_len, self.max_position_embeddings):
set_cos_sin_cache(self,
seq_len=max_seq_len,
device=query.device,
dtype=torch.float32)
# bsnd/bnsd
if positions is not None:
cos = self.embed(positions, self.cos)
sin = self.embed(positions, self.sin)
self.cos_embed = cos
self.sin_embed = sin
else:
cos = self.cos_embed
sin = self.sin_embed
query = query.view(*query.shape[:-1], -1,
self.head_size).contiguous()
key = key.view(*key.shape[:-1], -1, self.head_size).contiguous()
cos = cos.unsqueeze(-2).unsqueeze(-2)
sin = sin.unsqueeze(-2).unsqueeze(-2)
query = query.unsqueeze(1)
key = key.unsqueeze(1)
q_embed, k_embed = torch_npu.npu_apply_rotary_pos_emb(
query, key, cos, sin)
return q_embed.flatten(-2), k_embed.flatten(-2)
else:
return rope_forward_oot(self, positions, query, key, offsets,
is_neox_style_override,
is_qwen_torchair) # type: ignore
RotaryEmbedding.__init__ = qwen_rope_init_func
RotaryEmbedding.forward_oot = rope_forward
class AscendDeepseekScalingRotaryEmbedding(DeepseekScalingRotaryEmbedding):
# Note: we adopt the native huggingface deepseek rope initialization code from
# https://huggingface.co/deepseek-ai/DeepSeek-V3-0324/blob/main/modeling_deepseek.py for
# its more ascend compute friendly
DeepseekScalingRotaryEmbedding.__init__ = deepseek_rope_init_func
DeepseekScalingRotaryEmbedding.forward = native_rope_deepseek_forward
def __init__(
self,
head_size: int,
rotary_dim: int,
max_position_embeddings: int,
base: int,
is_neox_style: bool,
scaling_factor: float,
dtype: torch.dtype,
*,
extrapolation_factor: float = 1,
attn_factor: float = 1,
beta_fast: int = 32,
beta_slow: int = 1,
mscale: float = 1,
mscale_all_dim: float = 0,
) -> None:
# Note: we adopt the native huggingface deepseek rope initialization code from
# https://huggingface.co/deepseek-ai/DeepSeek-V3-0324/blob/main/modeling_deepseek.py for
# its more ascend compute friendly
self.scaling_factor = scaling_factor
self.extrapolation_factor = extrapolation_factor
self.attn_factor = attn_factor
self.beta_fast = beta_fast
self.beta_slow = beta_slow
# Get n-d magnitude scaling corrected for interpolation.
self.mscale = float(
self._yarn_get_mscale(self.scaling_factor, float(mscale)) /
self._yarn_get_mscale(self.scaling_factor, float(mscale_all_dim)) *
attn_factor)
super(DeepseekScalingRotaryEmbedding,
self).__init__(head_size, rotary_dim, max_position_embeddings,
base, is_neox_style, dtype)
self.max_seq_len = max_position_embeddings
self._set_cos_sin_cache(seq_len=max_position_embeddings,
device=NPUPlatform.device_type,
dtype=dtype)
def _yarn_get_mscale(self, scale: float = 1, mscale: float = 1) -> float:
if scale <= 1:
return 1.0
return 0.1 * mscale * math.log(scale) + 1.0
def _rotate_half(self, x):
"""Rotates half the hidden dims of the input."""
x1 = x[..., :x.shape[-1] // 2]
x2 = x[..., x.shape[-1] // 2:]
return torch.cat((-x2, x1), dim=-1)
def _yarn_linear_ramp_mask(self, min_value, max_value, dim):
# Note: The if conditional branch is not used here
# to solve MTP compilation error.
max_value += (min_value == max_value).float() * 0.001
linear_func = (torch.arange(dim, dtype=torch.float32) -
min_value) / (max_value - min_value)
ramp_func = torch.clamp(linear_func, 0, 1)
return ramp_func
# Inverse dim formula to find dim based on number of rotations
def _yarn_find_correction_dim(self,
num_rotations,
dim,
base=10000,
max_position_embeddings=2048):
# Note: use torch instead of math to solve MTP compilation error.
return (dim * torch.log(
torch.tensor(max_position_embeddings) /
(num_rotations * 2 * torch.pi))) / (2 *
torch.log(torch.tensor(base)))
# Find dim range bounds based on rotations
def _yarn_find_correction_range(self,
low_rot,
high_rot,
dim,
base=10000,
max_position_embeddings=2048):
# Note: use torch instead of math to solve MTP compilation error.
low = torch.floor(
self._yarn_find_correction_dim(low_rot, dim, base,
max_position_embeddings))
high = torch.ceil(
self._yarn_find_correction_dim(high_rot, dim, base,
max_position_embeddings))
# Note: use torch instead of max/min to solve MTP compilation error.
return torch.clamp(low, min=0), torch.clamp(high, max=dim - 1)
# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
def _apply_rotary_pos_emb(self,
q,
k,
cos,
sin,
position_ids,
unsqueeze_dim=1):
"""Applies Rotary Position Embedding to the query and key tensors.
Args:
q (`torch.Tensor`): The query tensor.
k (`torch.Tensor`): The key tensor.
cos (`torch.Tensor`): The cosine part of the rotary embedding.
sin (`torch.Tensor`): The sine part of the rotary embedding.
position_ids (`torch.Tensor`):
The position indices of the tokens corresponding to the query and key tensors. For example, this can be
used to pass offsetted position ids when working with a KV-cache.
unsqueeze_dim (`int`, *optional*, defaults to 1):
The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
Returns:
`tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
"""
cos = cos[position_ids]
sin = sin[position_ids]
cos = cos[:, None, None, :]
sin = sin[:, None, None, :]
if len(q.shape) == 3:
q = q[:, :, None, :]
if len(k.shape) == 2:
k = k[:, None, None, :]
elif len(k.shape) == 3:
k = k[:, :, None, :]
b, h_q, s, d = q.shape
q = q.view(b, h_q, s, d // 2, 2).transpose(4, 3).reshape(b, h_q, s, d)
b, h_k, s, d = k.shape
k = k.view(b, h_k, s, d // 2, 2).transpose(4, 3).reshape(b, h_k, s, d)
q_embed = (q * cos) + (self._rotate_half(q) * sin)
k_embed = (k * cos) + (self._rotate_half(k) * sin)
q_embed = q_embed.view(b, h_q, d)
k_embed = k_embed.view(b, h_k, d)
return q_embed, k_embed
def _set_cos_sin_cache(self, seq_len, device, dtype):
self.max_seq_len_cached = seq_len
dim = self.rotary_dim
freq_extra = 1.0 / (self.base**(
torch.arange(0, dim, 2, dtype=torch.float32, device=device) / dim))
freq_inter = 1.0 / (self.scaling_factor * self.base**(
torch.arange(0, dim, 2, dtype=torch.float32, device=device) / dim))
low, high = self._yarn_find_correction_range(
self.beta_fast,
self.beta_slow,
dim,
self.base,
self.max_position_embeddings,
)
inv_freq_mask = 1.0 - self._yarn_linear_ramp_mask(
low, high, dim // 2).to(device=device, dtype=torch.float32)
inv_freq = freq_inter * (1 -
inv_freq_mask) + freq_extra * inv_freq_mask
self.register_buffer("inv_freq", inv_freq, persistent=False)
t = torch.arange(seq_len * self.scaling_factor,
device=device,
dtype=torch.float32)
freqs = torch.outer(t, inv_freq)
cos_cached = torch.cat([freqs, freqs], dim=-1).cos() * self.mscale
sin_cached = torch.cat([freqs, freqs], dim=-1).sin() * self.mscale
cos_cached = cos_cached.to(dtype)
sin_cached = sin_cached.to(dtype)
cache = torch.cat(
[freqs.cos() * self.mscale,
freqs.sin() * self.mscale], dim=-1).to(dtype)
self.register_buffer("cos_sin_cache", cache, persistent=False)
self.register_buffer("cos_cached", cos_cached, persistent=False)
self.register_buffer("sin_cached", sin_cached, persistent=False)
def forward(self,
positions: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
offsets: Optional[torch.Tensor] = None,
max_seq_len: Optional[int] = None):
if max_seq_len is not None and max_seq_len > self.max_seq_len:
self._set_cos_sin_cache(max_seq_len, query.device, query.dtype)
if len(key.shape) == 2:
key = key[:, None, :]
# Note: we implement the non neox_style method with shuffle the last dim and neox style
# calculation method which is also more compute friendly to the ascend machine
# https://huggingface.co/deepseek-ai/DeepSeek-V3-0324/blob/main/modeling_deepseek.py
neox_style = True
if self.is_neox_style is False:
b, h_q, d = query.shape
query = query.view(b, h_q, d // 2,
2).transpose(3, 2).reshape(b, h_q, d)
b, h_k, d = key.shape
key = key.view(b, h_k, d // 2, 2).transpose(3,
2).reshape(b, h_k, d)
q_pe, k_pe = rope_forward_oot(self, positions, query, key, offsets,
neox_style)
return q_pe, k_pe

7
vllm_ascend/utils.py
View File

@@ -478,9 +478,16 @@ def register_ascend_customop():
from vllm_ascend.ops.linear import (AscendMlpColumnParallelLinear,
AscendMlpMergedColumnParallelLinear,
AscendMlpRowParallelLinear)
from vllm_ascend.ops.rotary_embedding import (
AscendDeepseekScalingRotaryEmbedding, AscendRotaryEmbedding)
CustomOp.register_oot(_decorated_op_cls=AscendQuickGELU, name="QuickGELU")
CustomOp.register_oot(_decorated_op_cls=AscendSiluAndMul,
name="SiluAndMul")
CustomOp.register_oot(_decorated_op_cls=AscendRotaryEmbedding,
name="RotaryEmbedding")
CustomOp.register_oot(
_decorated_op_cls=AscendDeepseekScalingRotaryEmbedding,
name="DeepseekScalingRotaryEmbedding")
if envs_ascend.VLLM_ASCEND_ENABLE_MLP_OPTIMIZE:
CustomOp.register_oot(_decorated_op_cls=AscendMlpColumnParallelLinear,
name="ColumnParallelLinear")