add the local attention

.gitignore

@@ -135,3 +135,4 @@ experimental/*
 # Weights and Biases logs
 wandb/
 /examples/translation/iwslt14.tokenized.de-en/
+toy/

fairseq/models/speech_to_text/s2t_transformer.py

@@ -150,6 +150,7 @@ class S2TTransformerModel(FairseqEncoderDecoderModel):
                 "selfattn",
                 "rel_selfattn",
                 "relative",
+                "local",
             ],
             help="transformer encoder self-attention layer type"
         )
@@ -187,6 +188,7 @@ class S2TTransformerModel(FairseqEncoderDecoderModel):
                 "selfattn",
                 "rel_selfattn",
                 "relative",
+                "local",
             ],
             help="transformer decoder self-attention layer type"
         )
@@ -277,6 +279,29 @@ class S2TTransformerModel(FairseqEncoderDecoderModel):
             default="learnable_dense",
             help='decoder layer history type'
         )
+
+        parser.add_argument(
+            '--hard-mask-window',
+            type=float,
+            metavar="D",
+            default=0,
+            help='window size of local mask'
+        )
+        parser.add_argument(
+            '--gauss-mask-sigma',
+            type=float,
+            metavar="D",
+            default=0,
+            help='standard deviation of the gauss mask'
+        )
+        parser.add_argument(
+            '--init-mask-weight',
+            type=float,
+            metavar="D",
+            default=0.5,
+            help='initialized weight for local mask'
+        )
+
         pass
 
     @classmethod
@@ -627,6 +652,10 @@ def base_architecture(args):
     args.max_decoder_relative_length = getattr(args, 'max_decoder_relative_length', -1)
     args.k_only = getattr(args, 'k_only', True)
 
+    args.hard_mask_window = getattr(args, 'hard_mask_window', 0)
+    args.gauss_mask_sigma = getattr(args, 'gauss_mask_sigma', 0)
+    args.init_mask_weight = getattr(args, 'init_mask_weight', 0)
+
 
 @register_model_architecture("s2t_transformer", "s2t_transformer_s")
 def s2t_transformer_s(args):

fairseq/modules/__init__.py

@@ -26,6 +26,7 @@ from .layer_norm import Fp32LayerNorm, LayerNorm
 from .learned_positional_embedding import LearnedPositionalEmbedding
 from .lightweight_convolution import LightweightConv, LightweightConv1dTBC
 from .linearized_convolution import LinearizedConvolution
+from .local_multihead_attention import LocalMultiheadAttention
 from .multihead_attention import MultiheadAttention
 from .positional_embedding import PositionalEmbedding
 from .rel_position_multihead_attention import RelPositionMultiheadAttention
@@ -70,6 +71,7 @@ __all__ = [
     "LightweightConv1dTBC",
     "LightweightConv",
     "LinearizedConvolution",
+    "LocalMultiheadAttention",
     "MultiheadAttention",
     "PositionalEmbedding",
     "RelPositionMultiheadAttention",

fairseq/modules/local_multihead_attention.py

+# Copyright (c) Facebook, Inc. and its affiliates.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.
+
+import math
+from typing import Dict, Optional, Tuple
+
+import torch
+import torch.nn.functional as F
+from fairseq import utils
+from fairseq.incremental_decoding_utils import with_incremental_state
+from fairseq.modules.fairseq_dropout import FairseqDropout
+from fairseq.modules.quant_noise import quant_noise
+from torch import Tensor, nn
+from torch.nn import Parameter
+
+
+@with_incremental_state
+class LocalMultiheadAttention(nn.Module):
+    """Multi-headed attention.
+
+    See "Attention Is All You Need" for more details.
+    """
+
+    def __init__(
+        self,
+        embed_dim,
+        num_heads,
+        kdim=None,
+        vdim=None,
+        dropout=0.0,
+        bias=True,
+        add_bias_kv=False,
+        add_zero_attn=False,
+        self_attention=False,
+        encoder_decoder_attention=False,
+        q_noise=0.0,
+        qn_block_size=8,
+        hard_mask_window=0,
+        gauss_mask_sigma=0,
+        init_mask_weight=0.5,
+    ):
+        super().__init__()
+        self.embed_dim = embed_dim
+        self.kdim = kdim if kdim is not None else embed_dim
+        self.vdim = vdim if vdim is not None else embed_dim
+        self.qkv_same_dim = self.kdim == embed_dim and self.vdim == embed_dim
+
+        self.num_heads = num_heads
+        self.dropout_module = FairseqDropout(
+            dropout, module_name=self.__class__.__name__
+        )
+
+        self.head_dim = embed_dim // num_heads
+        assert (
+            self.head_dim * num_heads == self.embed_dim
+        ), "embed_dim must be divisible by num_heads"
+        self.scaling = self.head_dim ** -0.5
+
+        self.self_attention = self_attention
+        self.encoder_decoder_attention = encoder_decoder_attention
+
+        assert not self.self_attention or self.qkv_same_dim, (
+            "Self-attention requires query, key and " "value to be of the same size"
+        )
+
+        self.k_proj = quant_noise(
+            nn.Linear(self.kdim, embed_dim, bias=bias), q_noise, qn_block_size
+        )
+        self.v_proj = quant_noise(
+            nn.Linear(self.vdim, embed_dim, bias=bias), q_noise, qn_block_size
+        )
+        self.q_proj = quant_noise(
+            nn.Linear(embed_dim, embed_dim, bias=bias), q_noise, qn_block_size
+        )
+
+        self.out_proj = quant_noise(
+            nn.Linear(embed_dim, embed_dim, bias=bias), q_noise, qn_block_size
+        )
+
+        if add_bias_kv:
+            self.bias_k = Parameter(torch.Tensor(1, 1, embed_dim))
+            self.bias_v = Parameter(torch.Tensor(1, 1, embed_dim))
+        else:
+            self.bias_k = self.bias_v = None
+
+        self.add_zero_attn = add_zero_attn
+
+        self.reset_parameters()
+
+        self.onnx_trace = False
+
+        if hard_mask_window != 0:
+            self.hard_mask_window = hard_mask_window
+        else:
+            self.hard_mask_window = 0
+
+        if gauss_mask_sigma != 0:
+            self.multihead_gauss_mask_sigma = Parameter(torch.Tensor(num_heads, 1, 1))
+            nn.init.constant_(self.multihead_gauss_mask_sigma, gauss_mask_sigma)
+        else:
+            self.multihead_gauss_mask_sigma = None
+
+        self.multihead_mask_weight = Parameter(torch.Tensor(num_heads, 1, 1))
+        nn.init.constant_(self.multihead_mask_weight, init_mask_weight)
+
+    def prepare_for_onnx_export_(self):
+        self.onnx_trace = True
+
+    def reset_parameters(self):
+        if self.qkv_same_dim:
+            # Empirically observed the convergence to be much better with
+            # the scaled initialization
+            nn.init.xavier_uniform_(self.k_proj.weight, gain=1 / math.sqrt(2))
+            nn.init.xavier_uniform_(self.v_proj.weight, gain=1 / math.sqrt(2))
+            nn.init.xavier_uniform_(self.q_proj.weight, gain=1 / math.sqrt(2))
+        else:
+            nn.init.xavier_uniform_(self.k_proj.weight)
+            nn.init.xavier_uniform_(self.v_proj.weight)
+            nn.init.xavier_uniform_(self.q_proj.weight)
+
+        nn.init.xavier_uniform_(self.out_proj.weight)
+        if self.out_proj.bias is not None:
+            nn.init.constant_(self.out_proj.bias, 0.0)
+        if self.bias_k is not None:
+            nn.init.xavier_normal_(self.bias_k)
+        if self.bias_v is not None:
+            nn.init.xavier_normal_(self.bias_v)
+
+    def forward(
+        self,
+        query,
+        key: Optional[Tensor],
+        value: Optional[Tensor],
+        key_padding_mask: Optional[Tensor] = None,
+        incremental_state: Optional[Dict[str, Dict[str, Optional[Tensor]]]] = None,
+        need_weights: bool = True,
+        static_kv: bool = False,
+        attn_mask: Optional[Tensor] = None,
+        before_softmax: bool = False,
+        need_head_weights: bool = False,
+    ) -> Tuple[Tensor, Optional[Tensor]]:
+        """Input shape: Time x Batch x Channel
+
+        Args:
+            key_padding_mask (ByteTensor, optional): mask to exclude
+                keys that are pads, of shape `(batch, src_len)`, where
+                padding elements are indicated by 1s.
+            need_weights (bool, optional): return the attention weights,
+                averaged over heads (default: False).
+            attn_mask (ByteTensor, optional): typically used to
+                implement causal attention, where the mask prevents the
+                attention from looking forward in time (default: None).
+            before_softmax (bool, optional): return the raw attention
+                weights and values before the attention softmax.
+            need_head_weights (bool, optional): return the attention
+                weights for each head. Implies *need_weights*. Default:
+                return the average attention weights over all heads.
+        """
+        if need_head_weights:
+            need_weights = True
+
+        is_tpu = query.device.type == "xla"
+
+        tgt_len, bsz, embed_dim = query.size()
+        assert embed_dim == self.embed_dim
+        assert list(query.size()) == [tgt_len, bsz, embed_dim]
+
+        if incremental_state is not None:
+            saved_state = self._get_input_buffer(incremental_state)
+            if saved_state is not None and "prev_key" in saved_state:
+                # previous time steps are cached - no need to recompute
+                # key and value if they are static
+                if static_kv:
+                    assert self.encoder_decoder_attention and not self.self_attention
+                    key = value = None
+        else:
+            saved_state = None
+
+        if self.self_attention:
+            q = self.q_proj(query)
+            k = self.k_proj(query)
+            v = self.v_proj(query)
+        elif self.encoder_decoder_attention:
+            # encoder-decoder attention
+            q = self.q_proj(query)
+            if key is None:
+                assert value is None
+                k = v = None
+            else:
+                k = self.k_proj(key)
+                v = self.v_proj(key)
+
+        else:
+            assert key is not None and value is not None
+            q = self.q_proj(query)
+            k = self.k_proj(key)
+            v = self.v_proj(value)
+        q *= self.scaling
+
+        if self.bias_k is not None:
+            assert self.bias_v is not None
+            k = torch.cat([k, self.bias_k.repeat(1, bsz, 1)])
+            v = torch.cat([v, self.bias_v.repeat(1, bsz, 1)])
+            if attn_mask is not None:
+                attn_mask = torch.cat(
+                    [attn_mask, attn_mask.new_zeros(attn_mask.size(0), 1)], dim=1
+                )
+            if key_padding_mask is not None:
+                key_padding_mask = torch.cat(
+                    [
+                        key_padding_mask,
+                        key_padding_mask.new_zeros(key_padding_mask.size(0), 1),
+                    ],
+                    dim=1,
+                )
+
+        q = (
+            q.contiguous()
+            .view(tgt_len, bsz * self.num_heads, self.head_dim)
+            .transpose(0, 1)
+        )
+        if k is not None:
+            k = (
+                k.contiguous()
+                .view(-1, bsz * self.num_heads, self.head_dim)
+                .transpose(0, 1)
+            )
+        if v is not None:
+            v = (
+                v.contiguous()
+                .view(-1, bsz * self.num_heads, self.head_dim)
+                .transpose(0, 1)
+            )
+
+        if saved_state is not None:
+            # saved states are stored with shape (bsz, num_heads, seq_len, head_dim)
+            if "prev_key" in saved_state:
+                _prev_key = saved_state["prev_key"]
+                assert _prev_key is not None
+                prev_key = _prev_key.view(bsz * self.num_heads, -1, self.head_dim)
+                if static_kv:
+                    k = prev_key
+                else:
+                    assert k is not None
+                    k = torch.cat([prev_key, k], dim=1)
+            if "prev_value" in saved_state:
+                _prev_value = saved_state["prev_value"]
+                assert _prev_value is not None
+                prev_value = _prev_value.view(bsz * self.num_heads, -1, self.head_dim)
+                if static_kv:
+                    v = prev_value
+                else:
+                    assert v is not None
+                    v = torch.cat([prev_value, v], dim=1)
+            prev_key_padding_mask: Optional[Tensor] = None
+            if "prev_key_padding_mask" in saved_state:
+                prev_key_padding_mask = saved_state["prev_key_padding_mask"]
+            assert k is not None and v is not None
+            key_padding_mask = LocalMultiheadAttention._append_prev_key_padding_mask(
+                key_padding_mask=key_padding_mask,
+                prev_key_padding_mask=prev_key_padding_mask,
+                batch_size=bsz,
+                src_len=k.size(1),
+                static_kv=static_kv,
+            )
+
+            saved_state["prev_key"] = k.view(bsz, self.num_heads, -1, self.head_dim)
+            saved_state["prev_value"] = v.view(bsz, self.num_heads, -1, self.head_dim)
+            saved_state["prev_key_padding_mask"] = key_padding_mask
+            # In this branch incremental_state is never None
+            assert incremental_state is not None
+            incremental_state = self._set_input_buffer(incremental_state, saved_state)
+        assert k is not None
+        src_len = k.size(1)
+
+        # This is part of a workaround to get around fork/join parallelism
+        # not supporting Optional types.
+        if key_padding_mask is not None and key_padding_mask.dim() == 0:
+            key_padding_mask = None
+
+        if key_padding_mask is not None:
+            assert key_padding_mask.size(0) == bsz
+            assert key_padding_mask.size(1) == src_len
+
+        if self.add_zero_attn:
+            assert v is not None
+            src_len += 1
+            k = torch.cat([k, k.new_zeros((k.size(0), 1) + k.size()[2:])], dim=1)
+            v = torch.cat([v, v.new_zeros((v.size(0), 1) + v.size()[2:])], dim=1)
+            if attn_mask is not None:
+                attn_mask = torch.cat(
+                    [attn_mask, attn_mask.new_zeros(attn_mask.size(0), 1)], dim=1
+                )
+            if key_padding_mask is not None:
+                key_padding_mask = torch.cat(
+                    [
+                        key_padding_mask,
+                        torch.zeros(key_padding_mask.size(0), 1).type_as(
+                            key_padding_mask
+                        ),
+                    ],
+                    dim=1,
+                )
+
+        attn_weights = torch.bmm(q, k.transpose(1, 2))
+        attn_weights = self.apply_sparse_mask(attn_weights, tgt_len, src_len, bsz)
+
+        assert list(attn_weights.size()) == [bsz * self.num_heads, tgt_len, src_len]
+
+        if self.hard_mask_window != 0:
+
+            hard_mask_window = self.hard_mask_window
+            if 0 < self.hard_mask_window <= 1:
+                hard_mask_window = int(src_len * self.hard_mask_window)
+
+            x1 = torch.arange(-1, src_len - 1, 1).view(-1, 1)
+            x2 = torch.arange(-1, src_len - 1, 1).view(1, -1)
+            dis = x2 - x1
+            mask_diag = torch.abs(dis) > hard_mask_window
+            mask_diag = mask_diag.unsqueeze(0)
+            attn_weights = attn_weights.masked_fill(mask_diag, float("-inf"))
+
+        if self.multihead_gauss_mask_sigma is not None:
+            x1 = torch.arange(-1, src_len - 1, 1).view(-1, 1)
+            x2 = torch.arange(-1, src_len - 1, 1).view(1, -1)
+            diag_growing = -(x1 - x2) ** 2 / 2.0
+            e_diag_gauss_mask = diag_growing.unsqueeze(0).repeat(self.num_heads, 1, 1)
+
+            e_sigma_square = 1 / torch.square(self.multihead_gauss_mask_sigma)
+            e_diag_gauss_mask_final = e_diag_gauss_mask * e_sigma_square
+
+            e_diag_gauss_mask_final = torch.unsqueeze(e_diag_gauss_mask_final, 0)
+            attn_weights = attn_weights.view(bsz, self.num_heads, tgt_len, src_len)
+            multihead_mask_weight = torch.sigmoid(self.multihead_mask_weight.unsqueeze(0))
+            attn_weights = (1 - multihead_mask_weight) * attn_weights + multihead_mask_weight * e_diag_gauss_mask_final
+            attn_weights = attn_weights.view(bsz * self.num_heads, tgt_len, src_len)
+
+        if attn_mask is not None:
+            attn_mask = attn_mask.unsqueeze(0)
+            if self.onnx_trace:
+                attn_mask = attn_mask.repeat(attn_weights.size(0), 1, 1)
+            attn_weights += attn_mask
+
+        if key_padding_mask is not None:
+            # don't attend to padding symbols
+            attn_weights = attn_weights.view(bsz, self.num_heads, tgt_len, src_len)
+            if not is_tpu:
+                attn_weights = attn_weights.masked_fill(
+                    key_padding_mask.unsqueeze(1).unsqueeze(2).to(torch.bool),
+                    float("-inf"),
+                )
+            else:
+                attn_weights = attn_weights.transpose(0, 2)
+                attn_weights = attn_weights.masked_fill(key_padding_mask, float("-inf"))
+                attn_weights = attn_weights.transpose(0, 2)
+            attn_weights = attn_weights.view(bsz * self.num_heads, tgt_len, src_len)
+
+        if before_softmax:
+            return attn_weights, v
+
+        attn_weights_float = utils.softmax(
+            attn_weights, dim=-1, onnx_trace=self.onnx_trace
+        )
+        attn_weights = attn_weights_float.type_as(attn_weights)
+        attn_probs = self.dropout_module(attn_weights)
+
+        assert v is not None
+        attn = torch.bmm(attn_probs, v)
+        assert list(attn.size()) == [bsz * self.num_heads, tgt_len, self.head_dim]
+        if self.onnx_trace and attn.size(1) == 1:
+            # when ONNX tracing a single decoder step (sequence length == 1)
+            # the transpose is a no-op copy before view, thus unnecessary
+            attn = attn.contiguous().view(tgt_len, bsz, embed_dim)
+        else:
+            attn = attn.transpose(0, 1).contiguous().view(tgt_len, bsz, embed_dim)
+        attn = self.out_proj(attn)
+        attn_weights: Optional[Tensor] = None
+        if need_weights:
+            attn_weights = attn_weights_float.view(
+                bsz, self.num_heads, tgt_len, src_len
+            ).transpose(1, 0)
+            if not need_head_weights:
+                # average attention weights over heads
+                attn_weights = attn_weights.mean(dim=0)
+
+        return attn, attn_weights
+
+    @staticmethod
+    def _append_prev_key_padding_mask(
+        key_padding_mask: Optional[Tensor],
+        prev_key_padding_mask: Optional[Tensor],
+        batch_size: int,
+        src_len: int,
+        static_kv: bool,
+    ) -> Optional[Tensor]:
+        # saved key padding masks have shape (bsz, seq_len)
+        if prev_key_padding_mask is not None and static_kv:
+            new_key_padding_mask = prev_key_padding_mask
+        elif prev_key_padding_mask is not None and key_padding_mask is not None:
+            new_key_padding_mask = torch.cat(
+                [prev_key_padding_mask.float(), key_padding_mask.float()], dim=1
+            )
+        # During incremental decoding, as the padding token enters and
+        # leaves the frame, there will be a time when prev or current
+        # is None
+        elif prev_key_padding_mask is not None:
+            filler = torch.zeros(
+                (batch_size, src_len - prev_key_padding_mask.size(1)),
+                device=prev_key_padding_mask.device,
+            )
+            new_key_padding_mask = torch.cat(
+                [prev_key_padding_mask.float(), filler.float()], dim=1
+            )
+        elif key_padding_mask is not None:
+            filler = torch.zeros(
+                (batch_size, src_len - key_padding_mask.size(1)),
+                device=key_padding_mask.device,
+            )
+            new_key_padding_mask = torch.cat(
+                [filler.float(), key_padding_mask.float()], dim=1
+            )
+        else:
+            new_key_padding_mask = prev_key_padding_mask
+        return new_key_padding_mask
+
+    @torch.jit.export
+    def reorder_incremental_state(
+        self,
+        incremental_state: Dict[str, Dict[str, Optional[Tensor]]],
+        new_order: Tensor,
+    ):
+        """Reorder buffered internal state (for incremental generation)."""
+        input_buffer = self._get_input_buffer(incremental_state)
+        if input_buffer is not None:
+            for k in input_buffer.keys():
+                input_buffer_k = input_buffer[k]
+                if input_buffer_k is not None:
+                    if self.encoder_decoder_attention and input_buffer_k.size(
+                        0
+                    ) == new_order.size(0):
+                        break
+                    input_buffer[k] = input_buffer_k.index_select(0, new_order)
+            incremental_state = self._set_input_buffer(incremental_state, input_buffer)
+        return incremental_state
+
+    def _get_input_buffer(
+        self, incremental_state: Optional[Dict[str, Dict[str, Optional[Tensor]]]]
+    ) -> Dict[str, Optional[Tensor]]:
+        result = self.get_incremental_state(incremental_state, "attn_state")
+        if result is not None:
+            return result
+        else:
+            empty_result: Dict[str, Optional[Tensor]] = {}
+            return empty_result
+
+    def _set_input_buffer(
+        self,
+        incremental_state: Dict[str, Dict[str, Optional[Tensor]]],
+        buffer: Dict[str, Optional[Tensor]],
+    ):
+        return self.set_incremental_state(incremental_state, "attn_state", buffer)
+
+    def apply_sparse_mask(self, attn_weights, tgt_len: int, src_len: int, bsz: int):
+        return attn_weights
+
+    def upgrade_state_dict_named(self, state_dict, name):
+        prefix = name + "." if name != "" else ""
+        items_to_add = {}
+        keys_to_remove = []
+        for k in state_dict.keys():
+            if k.endswith(prefix + "in_proj_weight"):
+                # in_proj_weight used to be q + k + v with same dimensions
+                dim = int(state_dict[k].shape[0] / 3)
+                items_to_add[prefix + "q_proj.weight"] = state_dict[k][:dim]
+                items_to_add[prefix + "k_proj.weight"] = state_dict[k][dim : 2 * dim]
+                items_to_add[prefix + "v_proj.weight"] = state_dict[k][2 * dim :]
+
+                keys_to_remove.append(k)
+
+                k_bias = prefix + "in_proj_bias"
+                if k_bias in state_dict.keys():
+                    dim = int(state_dict[k].shape[0] / 3)
+                    items_to_add[prefix + "q_proj.bias"] = state_dict[k_bias][:dim]
+                    items_to_add[prefix + "k_proj.bias"] = state_dict[k_bias][
+                        dim : 2 * dim
+                    ]
+                    items_to_add[prefix + "v_proj.bias"] = state_dict[k_bias][2 * dim :]
+
+                    keys_to_remove.append(prefix + "in_proj_bias")
+
+        for k in keys_to_remove:
+            del state_dict[k]
+
+        for key, value in items_to_add.items():
+            state_dict[key] = value

fairseq/modules/transformer_layer.py

@@ -12,7 +12,8 @@ from fairseq.modules import (
     LayerNorm,
     MultiheadAttention,
     RelPositionMultiheadAttention,
-    RelativeMultiheadAttention
+    RelativeMultiheadAttention,
+    LocalMultiheadAttention,
 )
 from fairseq.modules.fairseq_dropout import FairseqDropout
 from fairseq.modules.quant_noise import quant_noise
@@ -103,6 +104,21 @@ class TransformerEncoderLayer(nn.Module):
             else:
                 print("The maximum encoder relative length %d can not be -1!" % max_relative_length)
                 exit(1)
+        elif self.attn_type == "local":
+            hard_mask_window = getattr(args, "hard_mask_window", 0)
+            gauss_mask_sigma = getattr(args, "gauss_mask_sigma", 0)
+            init_mask_weight = getattr(args, "init_mask_weight", 0)
+            return LocalMultiheadAttention(
+                embed_dim,
+                args.encoder_attention_heads,
+                dropout=args.attention_dropout,
+                self_attention=True,
+                q_noise=self.quant_noise,
+                qn_block_size=self.quant_noise_block_size,
+                hard_mask_window=hard_mask_window,
+                gauss_mask_sigma=gauss_mask_sigma,
+                init_mask_weight=init_mask_weight
+            )
         else:
             print("The encoder attention type %s is not supported!" % self.attn_type)
             exit(1)