Refactor Attention (#1840)

This commit is contained in:
Woosuk Kwon
2023-11-29 15:37:31 -08:00
committed by GitHub
parent 0229c386c5
commit a9e4574261
16 changed files with 354 additions and 492 deletions
+185 -355
View File
@@ -1,5 +1,5 @@
"""Multi-head attention."""
from typing import Any, Dict, List, Optional
from typing import List, Optional
import torch
import torch.nn as nn
@@ -10,7 +10,6 @@ from xformers.ops.fmha.attn_bias import (BlockDiagonalCausalMask,
from vllm._C import ops
from vllm._C import cache_ops
from vllm.model_executor.input_metadata import InputMetadata
from vllm.model_executor.layers.rotary_embedding import get_rope
_SUPPORTED_HEAD_SIZES = [64, 80, 96, 112, 128, 256]
# Should be the same as PARTITION_SIZE in `paged_attention_v2_launcher`.
@@ -18,37 +17,39 @@ _PARTITION_SIZE = 512
class PagedAttention(nn.Module):
"""GPT-style multi-head PagedAttention.
"""MHA/MQA/GQA layer with PagedAttention.
This class takes query, key, and value tensors as input. The input tensors
can either contain prompt tokens or generation tokens, in addition to
paddings.
can either contain prompt tokens or generation tokens.
The class does the following:
1. Perform multi_query_kv_attention for the prompts. This operation does
not use the KV cache.
2. Wait for the cache operations (e.g., swap, copy) to finish. The cache
1. Wait for the cache operations (e.g., swap, copy) to finish. The cache
operations are issued by the cache engine before executing the forward
pass of the model, and they are executed asynchronously.
3. Reshape and store the input key and value tensors in the KV cache.
4. Perform single_query_cached_kv_attention for the generation tokens.
This operation reads the previous key and value tensors from the KV
cache.
5. Return the output tensor.
2. Reshape and store the input key and value tensors in the KV cache.
3. Perform (multi-head/multi-query/grouped-query) attention using either
xformers or the PagedAttention custom op.
4. Return the output tensor.
"""
def __init__(self,
num_heads: int,
head_size: int,
scale: float,
num_kv_heads: Optional[int] = None,
sliding_window: Optional[int] = None) -> None:
def __init__(
self,
num_heads: int,
head_size: int,
scale: float,
num_kv_heads: Optional[int] = None,
alibi_slopes: Optional[List[float]] = None,
sliding_window: Optional[int] = None,
) -> None:
super().__init__()
self.num_heads = num_heads
self.head_size = head_size
self.scale = float(scale)
self.num_kv_heads = num_heads if num_kv_heads is None else num_kv_heads
self.sliding_window = sliding_window
if alibi_slopes is not None:
alibi_slopes = torch.tensor(alibi_slopes, dtype=torch.float32)
self.register_buffer("alibi_slopes", alibi_slopes, persistent=False)
assert self.num_heads % self.num_kv_heads == 0
self.num_queries_per_kv = self.num_heads // self.num_kv_heads
@@ -60,153 +61,6 @@ class PagedAttention(nn.Module):
raise ValueError(f"head_size ({self.head_size}) is not supported. "
f"Supported head sizes: {_SUPPORTED_HEAD_SIZES}.")
def set_attn_bias(
self,
input_metadata: InputMetadata,
dtype: torch.dtype,
) -> None:
del dtype # Unused.
if input_metadata.attn_bias is not None:
# Already set by a previous layer.
return
prompt_lens = [input_metadata.max_prompt_len
] * input_metadata.num_prompts
attn_bias = BlockDiagonalCausalMask.from_seqlens(prompt_lens)
if self.sliding_window is not None:
attn_bias = attn_bias.make_local_attention(self.sliding_window)
input_metadata.attn_bias = attn_bias
def multi_query_kv_attention(
self,
output: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
value: torch.Tensor,
input_metadata: InputMetadata,
) -> torch.Tensor:
"""Normal attention for the prompt tokens.
Args:
output: shape = [num_prompt_tokens, num_heads, head_size]
query: shape = [num_prompt_tokens, num_heads, head_size]
key: shape = [num_prompt_tokens, num_kv_heads, head_size]
value: shape = [num_prompt_tokens, num_kv_heads, head_size]
input_metadata: metadata for paged attention.
"""
if self.num_kv_heads != self.num_heads:
# Project the key and value tensors to the desired number of heads.
query = query.view(query.shape[0], self.num_kv_heads,
self.num_queries_per_kv, query.shape[-1])
key = key[:, :,
None, :].expand(key.shape[0], self.num_kv_heads,
self.num_queries_per_kv, key.shape[-1])
value = value[:, :,
None, :].expand(value.shape[0], self.num_kv_heads,
self.num_queries_per_kv,
value.shape[-1])
# TODO(woosuk): The unsqueeze op may incur some CPU overhead. Optimize.
out = xops.memory_efficient_attention_forward(
query.unsqueeze(0),
key.unsqueeze(0),
value.unsqueeze(0),
attn_bias=input_metadata.attn_bias,
p=0.0,
scale=self.scale,
)
# TODO(woosuk): Unnecessary copy. Optimize.
output.copy_(out.view_as(output))
return output
def get_alibi_slopes(self) -> Optional[torch.Tensor]:
"""Returns the slopes for the alibi attention bias.
Returns:
slopes: shape = [num_heads]
"""
return None
def single_query_cached_kv_attention(
self,
output: torch.Tensor,
query: torch.Tensor,
key_cache: torch.Tensor,
value_cache: torch.Tensor,
input_metadata: InputMetadata,
alibi_slopes: Optional[torch.Tensor],
) -> None:
"""PagedAttention for the generation tokens.
Args:
output: shape = [num_generation_tokens, num_heads, head_size]
query: shape = [num_generation_tokens, num_heads, head_size]
key_cache: shape = [num_blocks, num_kv_heads, head_size/x,
block_size, x]
value_cache: shape = [num_blocks, num_kv_heads, head_size,
block_size]
input_metadata: metadata for paged attention.
alibi_slopes: shape = [num_heads]
"""
block_size = value_cache.shape[3]
num_seqs, num_heads, head_size = query.shape
max_num_partitions = (
(input_metadata.max_context_len + _PARTITION_SIZE - 1) //
_PARTITION_SIZE)
# NOTE(woosuk): We use a simple heuristic to decide whether to use
# PagedAttention V1 or V2. If the number of partitions is 1, we use
# V1 to avoid the overhead of reduction. Also, if the number of
# sequences or heads is large, we use V1 since there is enough work
# to parallelize.
# TODO(woosuk): Tune this heuristic.
# For context len > 8192, use V2 kernel to avoid shared memory shortage.
use_v1 = input_metadata.max_context_len <= 8192 and (
max_num_partitions == 1 or num_seqs * num_heads > 512)
if use_v1:
# Run PagedAttention V1.
ops.paged_attention_v1(
output,
query,
key_cache,
value_cache,
self.head_mapping,
self.scale,
input_metadata.block_tables,
input_metadata.context_lens,
block_size,
input_metadata.max_context_len,
alibi_slopes,
)
else:
# Run PagedAttention V2.
assert _PARTITION_SIZE % block_size == 0
tmp_output = torch.empty(
size=(num_seqs, num_heads, max_num_partitions, head_size),
dtype=output.dtype,
device=output.device,
)
exp_sums = torch.empty(
size=(num_seqs, num_heads, max_num_partitions),
dtype=torch.float32,
device=output.device,
)
max_logits = torch.empty_like(exp_sums)
ops.paged_attention_v2(
output,
exp_sums,
max_logits,
tmp_output,
query,
key_cache,
value_cache,
self.head_mapping,
self.scale,
input_metadata.block_tables,
input_metadata.context_lens,
block_size,
input_metadata.max_context_len,
alibi_slopes,
)
def forward(
self,
query: torch.Tensor,
@@ -219,9 +73,6 @@ class PagedAttention(nn.Module):
) -> torch.Tensor:
"""PagedAttention forward pass.
NOTE: The query, key, and value tensors must be sliced from a qkv
tensor of shape [batch_size, seq_len, 3 * num_heads * head_size].
Args:
query: shape = [batch_size, seq_len, num_heads * head_size]
key: shape = [batch_size, seq_len, num_kv_heads * head_size]
@@ -230,46 +81,28 @@ class PagedAttention(nn.Module):
block_size, x]
value_cache: shape = [num_blocks, num_kv_heads, head_size,
block_size]
input_metadata: metadata for paged attention.
input_metadata: metadata for the inputs.
cache_event: event to wait for the cache operations to finish.
Returns:
shape = [batch_size, seq_len, num_heads * head_size]
"""
batch_size, seq_len, _ = query.shape
batch_size, seq_len, hidden_size = query.shape
# Reshape the query, key, and value tensors.
query = query.view(-1, self.num_heads, self.head_size)
key = key.view(-1, self.num_kv_heads, self.head_size)
value = value.view(-1, self.num_kv_heads, self.head_size)
slot_mapping = input_metadata.slot_mapping.flatten()
# Pre-allocate the output tensor.
output = torch.empty_like(query)
# Compute the attention op for prompts.
num_prompt_tokens = input_metadata.num_prompt_tokens
if num_prompt_tokens > 0:
# Prompt run.
assert input_metadata.num_generation_tokens == 0
self.set_attn_bias(input_metadata, dtype=query.dtype)
self.multi_query_kv_attention(
output,
query,
key,
value,
input_metadata,
)
# Wait until the cache op is done.
if cache_event is not None:
cache_event.wait()
# Reshape the keys and values and store them in the cache.
# When key_cache and value_cache are not provided, the new key
# and value vectors will not be cached.
# If key_cache and value_cache are not provided, the new key and value
# vectors will not be cached. This happens during the initial memory
# profiling run.
if key_cache is not None and value_cache is not None:
key_to_cache = key
value_to_cache = value
slot_mapping = input_metadata.slot_mapping.view(-1)
if input_metadata.to_cache is not None:
key_to_cache = key_to_cache[input_metadata.to_cache]
value_to_cache = value_to_cache[input_metadata.to_cache]
@@ -283,178 +116,175 @@ class PagedAttention(nn.Module):
slot_mapping,
)
if input_metadata.num_generation_tokens > 0:
is_prompt = len(input_metadata.prompt_lens) > 0
if is_prompt:
# Prompt run.
if self.num_kv_heads != self.num_heads:
# As of Nov 2023, xformers only supports MHA. For MQA/GQA,
# project the key and value tensors to the desired number of
# heads.
# TODO(woosuk): Use MQA/GQA kernels for higher performance.
query = query.view(query.shape[0], self.num_kv_heads,
self.num_queries_per_kv, query.shape[-1])
key = key[:, :,
None, :].expand(key.shape[0], self.num_kv_heads,
self.num_queries_per_kv,
key.shape[-1])
value = value[:, :, None, :].expand(value.shape[0],
self.num_kv_heads,
self.num_queries_per_kv,
value.shape[-1])
# Set attention bias if not provided. This typically happens at the
# very attention layer of every iteration.
# FIXME(woosuk): This is a hack.
if input_metadata.attn_bias is None:
if self.alibi_slopes is None:
attn_bias = BlockDiagonalCausalMask.from_seqlens(
[seq_len] * batch_size)
if self.sliding_window is not None:
attn_bias = attn_bias.make_local_attention(
self.sliding_window)
input_metadata.attn_bias = attn_bias
else:
input_metadata.attn_bias = _make_alibi_bias(
self.alibi_slopes, batch_size, seq_len, query.dtype)
# TODO(woosuk): Too many view operations. Let's try to reduce them
# in the future for code readability.
if self.alibi_slopes is None:
query = query.unsqueeze(0)
key = key.unsqueeze(0)
value = value.unsqueeze(0)
else:
query = query.unflatten(0, (batch_size, seq_len))
key = key.unflatten(0, (batch_size, seq_len))
value = value.unflatten(0, (batch_size, seq_len))
out = xops.memory_efficient_attention_forward(
query,
key,
value,
attn_bias=input_metadata.attn_bias,
p=0.0,
scale=self.scale,
)
output = out.view_as(query)
else:
# Decoding run.
assert input_metadata.num_prompt_tokens == 0
assert key_cache is not None and value_cache is not None, (
"key_cache and value_cache must be provided when "
"generating tokens.")
# Compute the attention op for generation tokens.
self.single_query_cached_kv_attention(output, query, key_cache,
value_cache, input_metadata,
self.get_alibi_slopes())
output = _paged_attention(
query,
key_cache,
value_cache,
input_metadata,
self.head_mapping,
self.scale,
self.alibi_slopes,
)
# Reshape the output tensor.
# NOTE(woosuk): The output tensor may include paddings.
return output.view(batch_size, seq_len,
self.num_heads * self.head_size)
return output.view(batch_size, seq_len, hidden_size)
class PagedAttentionWithRoPE(PagedAttention):
"""PagedAttention with rotary positional embedding."""
def _make_alibi_bias(
alibi_slopes: torch.Tensor,
batch_size: int,
seq_len: int,
dtype: torch.dtype,
) -> LowerTriangularMaskWithTensorBias:
bias = torch.arange(seq_len, dtype=dtype)
# NOTE(zhuohan): HF uses
# `bias = bias[None, :].repeat(prompt_len, 1)`
# here. We find that both biases give the same results, but
# the bias below more accurately follows the original ALiBi
# paper.
bias = bias[None, :] - bias[:, None]
bias = bias.to(alibi_slopes.device)
def __init__(
self,
num_heads: int,
head_size: int,
scale: float,
rotary_dim: int,
max_position: int = 8192,
base: int = 10000,
num_kv_heads: Optional[int] = None,
is_neox_style: bool = True,
rope_scaling: Optional[Dict[str, Any]] = None,
sliding_window: Optional[int] = None,
) -> None:
super().__init__(num_heads,
head_size,
scale,
num_kv_heads,
sliding_window=sliding_window)
self.rotary_emb = get_rope(head_size, rotary_dim, max_position, base,
is_neox_style, rope_scaling)
# When using custom attention bias, xformers requires the bias to
# be sliced from a tensor whose length is a multiple of 8.
padded_len = (seq_len + 7) // 8 * 8
bias = torch.empty(
batch_size,
alibi_slopes.shape[0],
seq_len,
padded_len,
device=alibi_slopes.device,
dtype=dtype,
)[:, :, :, :seq_len].copy_(bias)
bias.mul_(alibi_slopes[:, None, None])
attn_bias = LowerTriangularMaskWithTensorBias(bias)
return attn_bias
def forward(
self,
positions: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
value: torch.Tensor,
key_cache: torch.Tensor,
value_cache: torch.Tensor,
input_metadata: InputMetadata,
cache_event: Optional[torch.cuda.Event],
) -> torch.Tensor:
""" PagedAttention forward pass with rotary embedding.
Args:
positions: shape = [batch_size, seq_len]
query: shape = [batch_size, seq_len, num_heads * head_size]
key: shape = [batch_size, seq_len, num_kv_heads * head_size]
value: shape = [batch_size, seq_len, num_kv_heads * head_size]
key_cache: shape = [num_blocks, num_kv_heads, head_size/x,
block_size, x]
value_cache: shape = [num_blocks, num_kv_heads, head_size,
block_size]
input_metadata: metadata for paged attention.
cache_event: event to wait for the cache operations to finish.
def _paged_attention(
query: torch.Tensor,
key_cache: torch.Tensor,
value_cache: torch.Tensor,
input_metadata: InputMetadata,
head_mapping: torch.Tensor,
scale: float,
alibi_slopes: Optional[torch.Tensor],
) -> torch.Tensor:
output = torch.empty_like(query)
Returns:
shape = [batch_size, seq_len, num_heads * head_size]
"""
# Apply rotary embedding to the query and key before passing them
# to the attention op.
query, key = self.rotary_emb(positions, query, key)
return super().forward(
block_size = value_cache.shape[3]
num_seqs, num_heads, head_size = query.shape
max_num_partitions = (
(input_metadata.max_context_len + _PARTITION_SIZE - 1) //
_PARTITION_SIZE)
# NOTE(woosuk): We use a simple heuristic to decide whether to use
# PagedAttention V1 or V2. If the number of partitions is 1, we use
# V1 to avoid the overhead of reduction. Also, if the number of
# sequences or heads is large, we use V1 since there is enough work
# to parallelize.
# TODO(woosuk): Tune this heuristic.
# For context len > 8192, use V2 kernel to avoid shared memory shortage.
use_v1 = input_metadata.max_context_len <= 8192 and (
max_num_partitions == 1 or num_seqs * num_heads > 512)
if use_v1:
# Run PagedAttention V1.
ops.paged_attention_v1(
output,
query,
key,
value,
key_cache,
value_cache,
input_metadata,
cache_event,
head_mapping,
scale,
input_metadata.block_tables,
input_metadata.context_lens,
block_size,
input_metadata.max_context_len,
alibi_slopes,
)
class PagedAttentionWithALiBi(PagedAttention):
"""PagedAttention with ALiBi attention bias."""
def __init__(self,
num_heads: int,
head_size: int,
scale: float,
slopes: List[float],
num_kv_heads: Optional[int] = None) -> None:
super().__init__(num_heads, head_size, scale, num_kv_heads)
assert len(slopes) == num_heads
slopes = torch.tensor(slopes, dtype=torch.float32)
self.register_buffer("alibi_slopes", slopes, persistent=False)
def set_attn_bias(self, input_metadata: InputMetadata,
dtype: torch.dtype) -> None:
if input_metadata.attn_bias is not None:
# Already set by a previous layer.
return
# Generates ALiBi mask based on the max prompt length.
max_prompt_len = input_metadata.max_prompt_len
bias = torch.arange(max_prompt_len, dtype=dtype)
# NOTE(zhuohan): HF uses
# `bias = bias[None, :].repeat(prompt_len, 1)`
# here. We find that both biases give the same results, but
# the bias below more accurately follows the original ALiBi
# paper.
bias = bias[None, :] - bias[:, None]
bias = bias.to(self.alibi_slopes.device)
# When using custom attention bias, xformers requires the bias to
# be sliced from a tensor whose length is a multiple of 8.
padded_len = (max_prompt_len + 7) // 8 * 8
bias = torch.empty(
input_metadata.num_prompts,
self.num_heads,
max_prompt_len,
padded_len,
device=self.alibi_slopes.device,
dtype=dtype,
)[:, :, :, :max_prompt_len].copy_(bias)
bias.mul_(self.alibi_slopes[:, None, None])
attn_bias = LowerTriangularMaskWithTensorBias(bias)
input_metadata.attn_bias = attn_bias
def multi_query_kv_attention(
self,
output: torch.Tensor,
query: torch.Tensor,
key: torch.Tensor,
value: torch.Tensor,
input_metadata: InputMetadata,
) -> torch.Tensor:
"""Attention with ALiBi bias for the prompt tokens.
Args:
output: shape = [num_prompt_tokens, num_heads, head_size]
query: shape = [num_prompt_tokens, num_heads, head_size]
key: shape = [num_prompt_tokens, num_kv_heads, head_size]
value: shape = [num_prompt_tokens, num_kv_heads, head_size]
input_metadata: metadata for paged attention.
"""
if self.num_kv_heads != self.num_heads:
# Project the key and value tensors to the desired number of heads.
query = query.view(query.shape[0], self.num_kv_heads,
self.num_queries_per_kv, query.shape[-1])
key = key[:, :,
None, :].expand(key.shape[0], self.num_kv_heads,
self.num_queries_per_kv, key.shape[-1])
value = value[:, :,
None, :].expand(value.shape[0], self.num_kv_heads,
self.num_queries_per_kv,
value.shape[-1])
batch_size = input_metadata.num_prompts
seq_len = input_metadata.max_prompt_len
out = xops.memory_efficient_attention_forward(
query.view(batch_size, seq_len, self.num_heads, self.head_size),
key.view(batch_size, seq_len, self.num_heads, self.head_size),
value.view(batch_size, seq_len, self.num_heads, self.head_size),
attn_bias=input_metadata.attn_bias,
p=0.0,
scale=self.scale,
else:
# Run PagedAttention V2.
assert _PARTITION_SIZE % block_size == 0
tmp_output = torch.empty(
size=(num_seqs, num_heads, max_num_partitions, head_size),
dtype=output.dtype,
device=output.device,
)
# TODO(woosuk): Unnecessary copy. Optimize.
output.copy_(out.view_as(output))
return output
def get_alibi_slopes(self) -> Optional[torch.Tensor]:
return self.alibi_slopes
exp_sums = torch.empty(
size=(num_seqs, num_heads, max_num_partitions),
dtype=torch.float32,
device=output.device,
)
max_logits = torch.empty_like(exp_sums)
ops.paged_attention_v2(
output,
exp_sums,
max_logits,
tmp_output,
query,
key_cache,
value_cache,
head_mapping,
scale,
input_metadata.block_tables,
input_metadata.context_lens,
block_size,
input_metadata.max_context_len,
alibi_slopes,
)
return output
@@ -277,8 +277,8 @@ def get_rope(
rotary_dim: int,
max_position: int,
base: int,
is_neox_style: bool,
rope_scaling: Optional[Dict[str, Any]],
is_neox_style: bool = True,
rope_scaling: Optional[Dict[str, Any]] = None,
) -> RotaryEmbedding:
if rope_scaling is None:
rotary_emb = RotaryEmbedding(head_size, rotary_dim, max_position, base,