mha.py

# Copyright (c) 2023, Tri Dao.

""""
The implementation was adopted from
https://github.com/Dao-AILab/flash-attention/blob/43950dda456e095969d842fca7a73c5bfe3cecd0
and made modifications to
    - support QK normalization
    - make ALiBi run with MHA (needed to cast alibi slopes to fp32)
    - make ALiBi run on CPU
"""

import math
from functools import partial

import torch
import torch.nn as nn
from einops import rearrange, repeat

try:
    from flash_attn import (
        flash_attn_kvpacked_func,
        flash_attn_qkvpacked_func,
        flash_attn_varlen_kvpacked_func,
        flash_attn_varlen_qkvpacked_func,
        flash_attn_with_kvcache,
    )
except ImportError:
    flash_attn_varlen_qkvpacked_func, flash_attn_varlen_kvpacked_func = None, None
    flash_attn_qkvpacked_func, flash_attn_kvpacked_func = None, None
    flash_attn_with_kvcache = None

try:
    from flash_attn.ops.fused_dense import ColumnParallelLinear, FusedDense, RowParallelLinear
except ImportError:
    FusedDense, ColumnParallelLinear, RowParallelLinear = None, None, None

try:
    from flash_attn.layers.rotary import RotaryEmbedding
except ImportError:
    RotaryEmbedding = None


# From https://github.com/ofirpress/attention_with_linear_biases/blob/4b92f28a005ead2567abe2359f633e73e08f3833/fairseq/models/transformer.py#L742
def get_alibi_slopes(nheads):
    def get_slopes_power_of_2(nheads):
        start = 2 ** (-(2 ** -(math.log2(nheads) - 3)))
        ratio = start
        return [start * ratio**i for i in range(nheads)]

    if math.log2(nheads).is_integer():
        return get_slopes_power_of_2(nheads)
    else:
        closest_power_of_2 = 2 ** math.floor(math.log2(nheads))
        return (
            get_slopes_power_of_2(closest_power_of_2)
            + get_alibi_slopes(2 * closest_power_of_2)[0::2][: nheads - closest_power_of_2]
        )

class MultiHeadLayernorm(nn.Module):
    def __init__(self, head_dim, num_heads, eps=1e-05, shared_normalization=False):
        super().__init__()
        if shared_normalization:
            self._reduce_dims = (-2, -1)
        else:
            self._reduce_dims = (-1,)
        self.weight = nn.Parameter(torch.ones((num_heads, head_dim)))
        self.bias = nn.Parameter(torch.zeros((num_heads, head_dim)))
        self.eps = eps

    def forward(self, x):
        var, mean = torch.var_mean(x, dim=self._reduce_dims, keepdim=True)
        x = (x - mean) / torch.sqrt(var + self.eps)
        return self.weight * x + self.bias

class FlashSelfAttention(nn.Module):
    """Implement the scaled dot product attention with softmax.
    Arguments
    ---------
        softmax_scale: The temperature to use for the softmax attention.
                      (default: 1/sqrt(d_keys) where d_keys is computed at
                      runtime)
        attention_dropout: The dropout rate to apply to the attention
                           (default: 0.0)
    """

    def __init__(
        self,
        causal=False,
        softmax_scale=None,
        attention_dropout=0.0,
        window_size=(-1, -1),
        alibi_slopes=None,
        deterministic=False,
        qk_norm_kwargs=None,
    ):
        super().__init__()
        assert flash_attn_varlen_qkvpacked_func is not None, "FlashAttention is not installed"
        assert flash_attn_qkvpacked_func is not None, "FlashAttention is not installed"
        self.causal = causal
        self.softmax_scale = softmax_scale
        self.drop = nn.Dropout(attention_dropout)
        self.register_buffer("alibi_slopes", alibi_slopes, persistent=False)
        self.window_size = window_size
        self.deterministic = deterministic
        if qk_norm_kwargs is not None:
            self.qk_norm = True
            self.q_layernorm = MultiHeadLayernorm(**qk_norm_kwargs)
            self.k_layernorm = MultiHeadLayernorm(**qk_norm_kwargs)
        else:
            self.qk_norm = False
            self.q_layernorm = None
            self.k_layernorm = None

    def forward(self, qkv, causal=None, cu_seqlens=None, max_seqlen=None):
        """Implements the multihead softmax attention.
        Arguments
        ---------
            qkv: The tensor containing the query, key, and value.
                If cu_seqlens is None and max_seqlen is None, then qkv has shape (B, S, 3, H, D).
                If cu_seqlens is not None and max_seqlen is not None, then qkv has shape
                (total, 3, H, D), where total is the sum of the sequence lengths in the batch.
            causal: if passed, will override self.causal
            cu_seqlens: (batch_size + 1,), dtype torch.int32. The cumulative sequence lengths
                of the sequences in the batch, used to index into qkv.
            max_seqlen: int. Maximum sequence length in the batch.
        Returns:
        --------
            out: (total, H, D) if cu_seqlens is not None and max_seqlen is not None,
                else (B, S, H, D).
        """
        assert qkv.dtype in [torch.float16, torch.bfloat16]
        assert qkv.is_cuda
        if self.qk_norm:
            if cu_seqlens is None:
                assert qkv.shape[2] == 3
                q, k, v = qkv.unbind(2)
                q = self.q_layernorm(q)
                k = self.k_layernorm(k)
                qkv = torch.stack([q, k, v], dim=2)
            else:
                assert qkv.shape[1] == 3
                q, k, v = qkv.unbind(1)
                q = self.q_layernorm(q)
                k = self.k_layernorm(k)
                qkv = torch.stack([q, k, v], dim=1)
        causal = self.causal if causal is None else causal
        unpadded = cu_seqlens is not None
        if self.alibi_slopes is not None:
            self.alibi_slopes = self.alibi_slopes.to(torch.float32)
        if unpadded:
            assert cu_seqlens.dtype == torch.int32
            assert max_seqlen is not None
            assert isinstance(max_seqlen, int)
            return flash_attn_varlen_qkvpacked_func(
                qkv,
                cu_seqlens,
                max_seqlen,
                self.drop.p if self.training else 0.0,
                softmax_scale=self.softmax_scale,
                causal=causal,
                alibi_slopes=self.alibi_slopes,
                window_size=self.window_size,
                deterministic=self.deterministic,
            )
        else:
            return flash_attn_qkvpacked_func(
                qkv,
                self.drop.p if self.training else 0.0,
                softmax_scale=self.softmax_scale,
                causal=causal,
                alibi_slopes=self.alibi_slopes,
                window_size=self.window_size,
                deterministic=self.deterministic,
            )


class FlashCrossAttention(nn.Module):
    """Implement the scaled dot product attention with softmax.
    Arguments
    ---------
        softmax_scale: The temperature to use for the softmax attention.
                      (default: 1/sqrt(d_keys) where d_keys is computed at
                      runtime)
        attention_dropout: The dropout rate to apply to the attention
                           (default: 0.0)
    """

    def __init__(
        self,
        causal=False,
        softmax_scale=None,
        attention_dropout=0.0,
        alibi_slopes=None,
        window_size=(-1, -1),
        deterministic=False,
    ):
        super().__init__()
        assert flash_attn_varlen_kvpacked_func is not None, "FlashAttention is not installed"
        assert flash_attn_kvpacked_func is not None, "FlashAttention is not installed"
        self.causal = causal
        self.softmax_scale = softmax_scale
        self.drop = nn.Dropout(attention_dropout)
        self.register_buffer("alibi_slopes", alibi_slopes, persistent=False)
        self.window_size = window_size
        self.deterministic = deterministic

    def forward(
        self,
        q,
        kv,
        causal=None,
        cu_seqlens=None,
        max_seqlen=None,
        cu_seqlens_k=None,
        max_seqlen_k=None,
    ):
        """Implements the multihead softmax attention.
        Arguments
        ---------
            q: The tensor containing the query. (B, Sq, H, D)
            kv: The tensor containing the key and value. (B, Sk, 2, H_k, D)
            causal: if passed, will override self.causal
            cu_seqlens: (batch_size + 1,), dtype torch.int32. The cumulative sequence lengths
                of the sequences in the batch, used to index into q.
            max_seqlen: int. Maximum sequence length in the batch of q.
            cu_seqlens_k: (batch_size + 1,), dtype torch.int32. The cumulative sequence lengths
                of the sequences in the batch, used to index into kv.
            max_seqlen_k: int. Maximum sequence length in the batch of k and v.
        """
        assert q.dtype in [torch.float16, torch.bfloat16]
        assert q.is_cuda and kv.is_cuda
        causal = self.causal if causal is None else causal
        unpadded = cu_seqlens is not None
        if self.alibi_slopes is not None:
            self.alibi_slopes = self.alibi_slopes.to(torch.float32)
        if unpadded:
            assert cu_seqlens.dtype == torch.int32
            assert max_seqlen is not None
            assert isinstance(max_seqlen, int)
            assert cu_seqlens_k is not None
            assert cu_seqlens_k.dtype == torch.int32
            assert max_seqlen_k is not None
            assert isinstance(max_seqlen, int)
            return flash_attn_varlen_kvpacked_func(
                q,
                kv,
                cu_seqlens,
                cu_seqlens_k,
                max_seqlen,
                max_seqlen_k,
                self.drop.p if self.training else 0.0,
                softmax_scale=self.softmax_scale,
                causal=causal,
                alibi_slopes=self.alibi_slopes,
                window_size=self.window_size,
                deterministic=self.deterministic,
            )
        else:
            batch_size, seqlen_q = q.shape[0], q.shape[1]
            seqlen_k = kv.shape[1]
            assert kv.shape[0] == batch_size and kv.shape[4] == q.shape[3]
            return flash_attn_kvpacked_func(
                q,
                kv,
                self.drop.p if self.training else 0.0,
                causal=causal,
                softmax_scale=self.softmax_scale,
                alibi_slopes=self.alibi_slopes,
                window_size=self.window_size,
                deterministic=self.deterministic,
            )


class SelfAttention(nn.Module):
    """Implement the scaled dot product attention with softmax.
    Arguments
    ---------
        softmax_scale: The temperature to use for the softmax attention.
                      (default: 1/sqrt(d_keys) where d_keys is computed at
                      runtime)
        attention_dropout: The dropout rate to apply to the attention
                           (default: 0.0)
    """
    def __init__(self,
                 causal=False,
                 softmax_scale=None,
                 attention_dropout=0.0,
                 alibi_slopes=None,
                 qk_norm_kwargs=None,
        ):
        super().__init__()
        self.causal = causal
        self.softmax_scale = softmax_scale
        self.drop = nn.Dropout(attention_dropout)
        self.register_buffer('alibi_slopes', alibi_slopes, persistent=False)
        if alibi_slopes is not None:
            self.register_buffer('linear_biases', self._build_linear_biases(16), persistent=False)
        else:
            self.linear_biases = None
        if qk_norm_kwargs is not None:
            self.qk_norm = True
            self.q_layernorm = MultiHeadLayernorm(**qk_norm_kwargs)
            self.k_layernorm = MultiHeadLayernorm(**qk_norm_kwargs)
        else:
            self.qk_norm = False
            self.q_layernorm = None
            self.k_layernorm = None

    def _build_linear_biases(self, seqlen):
        context_position = torch.arange(seqlen, device=self.alibi_slopes.device)[:, None]
        memory_position = torch.arange(seqlen, device=self.alibi_slopes.device)[None, :]
        # distance tensor is of shape (seqlen, seqlen)
        distance = torch.abs(memory_position - context_position)
        # alibi tensor is of shape (1, H, seqlen, seqlen)
        linear_biases = (distance[None, ...] * self.alibi_slopes[:, None, None])[None, ...]
        return linear_biases

    def forward(self, qkv, causal=None, key_padding_mask=None):
        """Implements the multihead softmax attention.
        Arguments
        ---------
            qkv: The tensor containing the query, key, and value. (B, S, 3, H, D)
            causal: if passed, will override self.causal
            key_padding_mask: boolean mask to apply to the attention weights. True means to keep,
                False means to mask out. (B, S)
        """
        batch_size, seqlen = qkv.shape[0], qkv.shape[1]
        causal = self.causal if causal is None else causal
        q, k, v = qkv.unbind(dim=2)
        if self.qk_norm:
            q = self.q_layernorm(q)
            k = self.k_layernorm(k)
        softmax_scale = self.softmax_scale or 1.0 / math.sqrt(q.shape[-1])
        scores = torch.einsum("bthd,bshd->bhts", q, k * softmax_scale)
        if key_padding_mask is not None:
            padding_mask = torch.full(
                (batch_size, seqlen), -10000.0, dtype=scores.dtype, device=scores.device
            )
            padding_mask.masked_fill_(key_padding_mask, 0.0)
            # TD [2022-09-30]: Adding is faster than masked_fill_ (idk why, just better kernel I guess)
            scores = scores + rearrange(padding_mask, "b s -> b 1 1 s")
        if self.alibi_slopes is not None:
            if seqlen > self.linear_biases.shape[-1]:
                self.linear_biases = self._build_linear_biases(seqlen)
            cropped_biases = self.linear_biases[..., :seqlen, :seqlen]
            scores = scores - cropped_biases
        if causal:
            # "triu_tril_cuda_template" not implemented for 'BFloat16'
            # So we have to construct the mask in float
            causal_mask = torch.triu(
                torch.full((seqlen, seqlen), -10000.0, device=scores.device), 1
            )
            # TD [2022-09-30]: Adding is faster than masked_fill_ (idk why, just better kernel I guess)
            scores = scores + causal_mask.to(dtype=scores.dtype)
        attention = torch.softmax(scores, dim=-1, dtype=v.dtype)
        attention_drop = self.drop(attention)
        output = torch.einsum("bhts,bshd->bthd", attention_drop, v)
        return output


class CrossAttention(nn.Module):
    """Implement the scaled dot product attention with softmax.
    Arguments
    ---------
        softmax_scale: The temperature to use for the softmax attention.
                      (default: 1/sqrt(d_keys) where d_keys is computed at
                      runtime)
        attention_dropout: The dropout rate to apply to the attention
                           (default: 0.0)
    """

    def __init__(self, causal=False, softmax_scale=None, attention_dropout=0.0):
        super().__init__()
        self.causal = causal
        self.softmax_scale = softmax_scale
        self.drop = nn.Dropout(attention_dropout)

    def forward(self, q, kv, causal=None, key_padding_mask=None):
        """Implements the multihead softmax attention.
        Arguments
        ---------
            q: The tensor containing the query. (B, Sq, H, D)
            kv: The tensor containing the key and value. (B, Sk, 2, H_k, D)
            causal: if passed, will override self.causal
            key_padding_mask: boolean mask to apply to the attention weights. True means to keep,
                False means to mask out. (B, Sk)
        """
        batch_size, seqlen_q = q.shape[0], q.shape[1]
        causal = self.causal if causal is None else causal
        seqlen_k = kv.shape[1]
        assert kv.shape[0] == batch_size and kv.shape[4] == q.shape[3]
        if kv.shape[3] != q.shape[2]:  # MQA/GQA
            kv = repeat(kv, "... hkv d -> ... (hkv g) d", g=q.shape[2] // kv.shape[3])
        k, v = kv.unbind(dim=2)
        softmax_scale = self.softmax_scale or 1.0 / math.sqrt(q.shape[-1])
        scores = torch.einsum("bthd,bshd->bhts", q, k * softmax_scale)
        if key_padding_mask is not None:
            padding_mask = torch.full(
                (batch_size, seqlen_k), -10000.0, dtype=scores.dtype, device=scores.device
            )
            padding_mask.masked_fill_(key_padding_mask, 0.0)
            # TD [2022-09-30]: Adding is faster than masked_fill_ (idk why, just better kernel I guess)
            scores = scores + rearrange(padding_mask, "b s -> b 1 1 s")
        if causal:
            # causal mask needs to take into account the difference between seqlen_q and seqlen_k
            row_idx = rearrange(
                torch.arange(seqlen_q, device=q.device, dtype=torch.long), "s -> s 1"
            )
            col_idx = torch.arange(seqlen_k, device=kv.device, dtype=torch.long)
            sk = (
                seqlen_k
                if key_padding_mask is None
                else rearrange(key_padding_mask.sum(-1), "b -> b 1 1 1")
            )
            causal_mask = col_idx > row_idx + sk - seqlen_q
            scores = scores.masked_fill(causal_mask, -10000.0)
        attention = torch.softmax(scores, dim=-1, dtype=v.dtype)
        attention_drop = self.drop(attention)
        output = torch.einsum("bhts,bshd->bthd", attention_drop, v)
        return output


class LinearResidual(nn.Linear):
    """Wrap nn.Linear to return the residual as well. For compatibility with FusedDense."""

    def forward(self, input: torch.Tensor) -> torch.Tensor:
        return super().forward(input), input


def _update_kv_cache(kv, inference_params, layer_idx):
    """kv: (batch_size, seqlen, 2, nheads, head_dim) or (batch_size, 1, 2, nheads, head_dim)"""
    # Pre-allocate memory for key-values for inference.
    num_heads, head_dim = kv.shape[-2:]
    if layer_idx not in inference_params.key_value_memory_dict:
        kv_cache = torch.empty(
            inference_params.max_batch_size,
            inference_params.max_seqlen,
            2,
            num_heads,
            head_dim,
            dtype=kv.dtype,
            device=kv.device,
        )
        inference_params.key_value_memory_dict[layer_idx] = kv_cache
    else:
        kv_cache = inference_params.key_value_memory_dict[layer_idx]
    # Adjust key and value for inference
    batch_start = inference_params.batch_size_offset
    batch_end = batch_start + kv.shape[0]
    sequence_start = inference_params.seqlen_offset
    sequence_end = sequence_start + kv.shape[1]
    assert batch_end <= kv_cache.shape[0]
    assert sequence_end <= kv_cache.shape[1]
    assert kv_cache is not None
    kv_cache[batch_start:batch_end, sequence_start:sequence_end, ...] = kv
    return kv_cache[batch_start:batch_end, :sequence_end, ...]


class MHA(nn.Module):
    """Multi-head self-attention and cross-attention"""

    def __init__(
        self,
        embed_dim,
        num_heads,
        num_heads_kv=None,
        cross_attn=False,
        qkv_proj_bias=True,
        out_proj_bias=True,
        dropout=0.0,
        softmax_scale=None,
        causal=False,
        layer_idx=None,
        dwconv=False,
        rotary_emb_dim=0,
        rotary_emb_base=10000.0,
        rotary_emb_scale_base=None,
        rotary_emb_interleaved=False,
        use_alibi=False,
        window_size=(-1, -1),
        fused_bias_fc=False,
        use_flash_attn=False,
        return_residual=False,
        checkpointing=False,
        device=None,
        dtype=None,
        qk_norm=False,
        qk_norm_kwargs=None,
    ) -> None:
        """
        num_heads_kv: can be used to toggle MQA / GQA. If None, use num_heads.
        return_residual: whether to return the input x along with the output. This is for
            performance reason: for post-norm architecture, returning the input allows us
            to fuse the backward of nn.Linear with the residual connection.
        """
        if qk_norm and cross_attn:
            raise NotImplementedError('QK normalization is only implemented for self-attention.')
        if qk_norm:
            qk_norm_kwargs = qk_norm_kwargs if qk_norm_kwargs is not None else {}
            qk_norm_kwargs.update({'num_heads': num_heads, 'head_dim': embed_dim  // num_heads})
        factory_kwargs = {"device": device, "dtype": dtype}
        super().__init__()
        self.embed_dim = embed_dim
        self.cross_attn = cross_attn
        self.causal = causal
        self.layer_idx = layer_idx
        self.dwconv = dwconv
        self.rotary_emb_dim = rotary_emb_dim
        self.use_flash_attn = use_flash_attn
        self.return_residual = return_residual
        self.checkpointing = checkpointing
        if use_alibi:
            assert not cross_attn or use_flash_attn, "ALiBi code path requires self-attention or cross-attention with flash_attn"
            alibi_slopes = torch.tensor(get_alibi_slopes(num_heads), device=device)
        else:
            alibi_slopes = None
        if window_size != (-1, -1):
            assert use_flash_attn, "Local (sliding window) attention code path requires flash_attn"

        self.num_heads = num_heads
        self.num_heads_kv = num_heads_kv if num_heads_kv is not None else num_heads
        assert (
            self.num_heads % self.num_heads_kv == 0
        ), "num_heads must be divisible by num_heads_kv"
        assert self.embed_dim % num_heads == 0, "embed_dim must be divisible by num_heads"
        self.head_dim = self.embed_dim // num_heads
        qkv_dim = self.head_dim * (self.num_heads + 2 * self.num_heads_kv)
        kv_dim = 2 * self.head_dim * self.num_heads_kv

        if self.rotary_emb_dim > 0:
            assert not cross_attn, "MHA with rotary embedding does not support cross-attention yet"
            assert RotaryEmbedding is not None, "rotary_emb is not installed"
            self.rotary_emb = RotaryEmbedding(
                self.rotary_emb_dim,
                base=rotary_emb_base,
                scale_base=rotary_emb_scale_base,
                interleaved=rotary_emb_interleaved,
                device=device,
            )

        if fused_bias_fc and FusedDense is None:
            raise ImportError("fused_dense is not installed")
        linear_cls = nn.Linear if not fused_bias_fc else FusedDense
        linear_resid_cls = (
            LinearResidual if not fused_bias_fc else partial(FusedDense, return_residual=True)
        )
        wqkv_cls = linear_cls if not self.return_residual else linear_resid_cls
        inner_attn_cls = (
            partial(FlashSelfAttention, alibi_slopes=alibi_slopes, window_size=window_size, qk_norm_kwargs=qk_norm_kwargs)
            if use_flash_attn
            else partial(SelfAttention, alibi_slopes=alibi_slopes, qk_norm_kwargs=qk_norm_kwargs)
        )
        inner_cross_attn_cls = (
            partial(FlashCrossAttention, alibi_slopes=alibi_slopes, window_size=window_size)
            if use_flash_attn
            else CrossAttention
        )
        if not self.cross_attn:
            self.Wqkv = wqkv_cls(embed_dim, qkv_dim, bias=qkv_proj_bias, **factory_kwargs)
        else:
            self.Wq = linear_cls(embed_dim, embed_dim, bias=qkv_proj_bias, **factory_kwargs)
            self.Wkv = wqkv_cls(embed_dim, kv_dim, bias=qkv_proj_bias, **factory_kwargs)
        if self.dwconv:
            if self.num_heads_kv == self.num_heads:
                self.dwconv_qkv = nn.Conv1d(
                    qkv_dim, qkv_dim, kernel_size=3, padding=2, groups=qkv_dim
                )
            else:
                self.dwconv_q = nn.Conv1d(
                    embed_dim, embed_dim, kernel_size=3, padding=2, groups=embed_dim
                )
                self.dwconv_kv = nn.Conv1d(kv_dim, kv_dim, kernel_size=3, padding=2, groups=kv_dim)
        self.inner_attn = inner_attn_cls(
            causal=causal,
            softmax_scale=softmax_scale,
            attention_dropout=dropout,
        )
        self.inner_cross_attn = inner_cross_attn_cls(
            causal=causal, softmax_scale=softmax_scale, attention_dropout=dropout
        )
        self.out_proj = linear_cls(embed_dim, embed_dim, bias=out_proj_bias, **factory_kwargs)

    def allocate_inference_cache(self, batch_size, max_seqlen, dtype=None):
        dtype = self.out_proj.weight.dtype if dtype is None else dtype
        device = self.out_proj.weight.device
        return torch.empty(
            batch_size,
            max_seqlen,
            2,
            self.num_heads_kv,
            self.head_dim,
            dtype=dtype,
            device=device,
        )

    def _update_kv_cache(self, kv, inference_params):
        """kv: (batch_size, seqlen, 2, nheads, head_dim) or (batch_size, 1, 2, nheads, head_dim)"""
        assert not self.dwconv, "Generation does not support dwconv yet"
        assert self.layer_idx is not None, "Generation requires layer_idx in the constructor"
        return _update_kv_cache(kv, inference_params, self.layer_idx)

    def _apply_rotary_update_kvcache_attention(self, q, kv, inference_params):
        """
        Fast path that combine 3 steps: apply rotary to Q and K, update kv cache, and apply attention.
        q: (batch_size, seqlen_q, nheads, head_dim)
        kv: (batch_size, seqlen_k, 2, nheads_kv, head_dim)
        """
        assert inference_params is not None and inference_params.seqlen_offset > 0
        assert self.use_flash_attn
        if self.rotary_emb_dim > 0:
            assert self.rotary_emb.scale is None, "This code path does not support xPos"
            self.rotary_emb._update_cos_sin_cache(
                inference_params.max_seqlen, device=q.device, dtype=q.dtype
            )
            rotary_cos, rotary_sin = self.rotary_emb._cos_cached, self.rotary_emb._sin_cached
        else:
            rotary_cos, rotary_sin = None, None
        batch = q.shape[0]
        kv_cache = inference_params.key_value_memory_dict[self.layer_idx][:batch]
        cache_seqlens = (
            inference_params.lengths_per_sample[:batch]
            if inference_params.lengths_per_sample is not None
            else inference_params.seqlen_offset
        )
        alibi_slopes = getattr(self.inner_cross_attn, "alibi_slopes", None)
        context = flash_attn_with_kvcache(
            q,
            kv_cache[:, :, 0],
            kv_cache[:, :, 1],
            kv[:, :, 0],
            kv[:, :, 1],
            rotary_cos=rotary_cos,
            rotary_sin=rotary_sin,
            cache_seqlens=cache_seqlens,
            softmax_scale=self.inner_cross_attn.softmax_scale,
            causal=self.inner_cross_attn.causal,
            rotary_interleaved=self.rotary_emb.interleaved if self.rotary_emb_dim > 0 else False,
            alibi_slopes=alibi_slopes,
        )
        return context

    def _update_kvcache_attention(self, q, kv, inference_params):
        """Write kv to inference_params, then do attention"""
        if (
            inference_params.seqlen_offset == 0
            or flash_attn_with_kvcache is None
            or not self.use_flash_attn
        ):
            # TODO: this only uses seqlen_offset and not lengths_per_sample.
            kv = self._update_kv_cache(kv, inference_params)
            return self.inner_cross_attn(q, kv)
        else:
            batch = q.shape[0]
            kv_cache = inference_params.key_value_memory_dict[self.layer_idx][:batch]
            cache_seqlens = (
                inference_params.lengths_per_sample[:batch]
                if inference_params.lengths_per_sample is not None
                else inference_params.seqlen_offset
            )
            alibi_slopes = getattr(self.inner_cross_attn, "alibi_slopes", None)
            return flash_attn_with_kvcache(
                q,
                kv_cache[:, :, 0],
                kv_cache[:, :, 1],
                kv[:, :, 0],
                kv[:, :, 1],
                cache_seqlens=cache_seqlens,
                softmax_scale=self.inner_cross_attn.softmax_scale,
                causal=self.inner_cross_attn.causal,
                alibi_slopes=alibi_slopes,
            )

    def forward(
        self,
        x,
        x_kv=None,
        key_padding_mask=None,
        cu_seqlens=None,
        max_seqlen=None,
        mixer_subset=None,
        inference_params=None,
        **kwargs,
    ):
        """
        Arguments:
            x: (batch, seqlen, hidden_dim) (where hidden_dim = num heads * head dim) if
                cu_seqlens is None and max_seqlen is None, else (total, hidden_dim) where total
                is the is the sum of the sequence lengths in the batch.
            x_kv: (batch, seqlen, hidden_dim), only applicable for cross-attention. If None, use x.
            cu_seqlens: (batch_size + 1,), dtype torch.int32. The cumulative sequence lengths
                of the sequences in the batch, used to index into x. Only applicable when using
                FlashAttention.
            max_seqlen: int. Maximum sequence length in the batch.
            key_padding_mask: boolean mask, True means to keep, False means to mask out.
                (batch, seqlen). Only applicable when not using FlashAttention.
            mixer_subset: for cross-attention only. If not None, will take a subset of x
                before applying the query projection. Useful for e.g., ViT where we only care
                about the CLS token in the last layer.
            inference_params: for generation. Adapted from Megatron-LM (and Apex)
            https://github.com/NVIDIA/apex/blob/3ff1a10f72ec07067c4e44759442329804ac5162/apex/transformer/testing/standalone_transformer_lm.py#L470
        """
        if cu_seqlens is not None:
            assert max_seqlen is not None
            assert key_padding_mask is None
            assert self.use_flash_attn
            assert not self.dwconv
            assert self.rotary_emb_dim == 0
        if key_padding_mask is not None:
            assert cu_seqlens is None
            assert max_seqlen is None
            assert not self.use_flash_attn
        if inference_params is not None:
            assert key_padding_mask is None
            assert cu_seqlens is None and max_seqlen is None
            assert not self.dwconv

        kwargs = (
            {"cu_seqlens": cu_seqlens, "max_seqlen": max_seqlen, **kwargs}
            if self.use_flash_attn
            else {"key_padding_mask": key_padding_mask, **kwargs}
        )
        seqlen_offset = (
            0
            if inference_params is None
            else (
                inference_params.lengths_per_sample
                if inference_params.lengths_per_sample is not None
                else inference_params.seqlen_offset
            )
        )
        rotary_max_seqlen = inference_params.max_seqlen if inference_params is not None else None
        batch, seqlen = x.shape[:2]
        if not self.cross_attn and self.num_heads_kv == self.num_heads:
            assert x_kv is None and mixer_subset is None
            if not self.return_residual:
                qkv = self.Wqkv(x)
            else:
                qkv, x = self.Wqkv(x)
            if self.dwconv:
                qkv = rearrange(
                    self.dwconv_qkv(rearrange(qkv, "b s d -> b d s"))[..., :-2], "b d s -> b s d"
                ).contiguous()
            qkv = rearrange(qkv, "... (three h d) -> ... three h d", three=3, d=self.head_dim)
            if (
                inference_params is None
                or inference_params.seqlen_offset == 0
                or (self.rotary_emb_dim == 0 or self.rotary_emb_dim % 16 != 0)
                or not self.use_flash_attn
            ):
                if self.rotary_emb_dim > 0:
                    qkv = self.rotary_emb(
                        qkv, seqlen_offset=seqlen_offset, max_seqlen=rotary_max_seqlen
                    )
                if inference_params is None:
                    if not self.checkpointing:
                        context = self.inner_attn(qkv, **kwargs)
                    else:
                        context = torch.utils.checkpoint.checkpoint(self.inner_attn, qkv, use_reentrant=False, **kwargs)
                else:
                    context = self._update_kvcache_attention(
                        qkv[:, :, 0], qkv[:, :, 1:], inference_params
                    )
            else:
                context = self._apply_rotary_update_kvcache_attention(
                    qkv[:, :, 0], qkv[:, :, 1:], inference_params
                )
        else:
            if self.cross_attn:
                if not self.return_residual:
                    q = self.Wq(x if mixer_subset is None else x[:, mixer_subset])
                    kv = self.Wkv(x_kv if x_kv is not None else x)
                else:
                    if x_kv is not None:
                        kv, x_kv = self.Wkv(x_kv)
                    else:
                        kv, x = self.Wkv(x)
                    q = self.Wq(x if mixer_subset is None else x[:, mixer_subset])
            else:
                assert self.num_heads_kv != self.num_heads
                if not self.return_residual:
                    qkv = self.Wqkv(x)
                else:
                    qkv, x = self.Wqkv(x)
                q = qkv[..., : self.num_heads * self.head_dim]
                kv = qkv[..., self.num_heads * self.head_dim :]
            q = rearrange(q, "... (h d) -> ... h d", d=self.head_dim)
            kv = rearrange(kv, "... (two hkv d) -> ... two hkv d", two=2, d=self.head_dim)
            if self.dwconv:
                q = rearrange(
                    self.dwconv_q(rearrange(q, "b s d -> b d s"))[..., :-2], "b d s -> b s d"
                ).contiguous()
                kv = rearrange(
                    self.dwconv_kv(rearrange(kv, "b s d -> b d s"))[..., :-2], "b d s -> b s d"
                ).contiguous()
            if (
                inference_params is None
                or inference_params.seqlen_offset == 0
                or (self.rotary_emb_dim == 0 or self.rotary_emb_dim % 16 != 0)
                or not self.use_flash_attn
            ):
                if self.rotary_emb_dim > 0:
                    q, kv = self.rotary_emb(
                        q, kv, seqlen_offset=seqlen_offset, max_seqlen=rotary_max_seqlen
                    )
                if inference_params is None:
                    if not self.checkpointing:
                        context = self.inner_cross_attn(q, kv, **kwargs)
                    else:
                        context = torch.utils.checkpoint.checkpoint(
                            self.inner_cross_attn, q, kv, use_reentrant=False, **kwargs
                        )
                else:
                    context = self._update_kvcache_attention(q, kv, inference_params)
            else:
                context = self._apply_rotary_update_kvcache_attention(q, kv, inference_params)
        out = self.out_proj(rearrange(context, "... h d -> ... (h d)"))
        return out if not self.return_residual else (out, x)