import math import torch from typing import Optional, Tuple from torch import nn from torch.nn import Parameter, Linear from torch.cuda.amp import autocast from modules.commons.layers import LayerNorm, Embedding from modules.commons.transformer import TransformerFFNLayer, MultiheadAttention from utils.nn.seq_utils import get_incremental_state, set_incremental_state, softmax, make_positions import torch.nn.functional as F DEFAULT_MAX_SOURCE_POSITIONS = 3000 DEFAULT_MAX_TARGET_POSITIONS = 3000 class SinusoidalPositionalEmbedding(nn.Module): """This module produces sinusoidal positional embeddings of any length. Padding symbols are ignored. """ def __init__(self, embedding_dim, padding_idx, init_size=1024): super().__init__() self.embedding_dim = embedding_dim self.padding_idx = padding_idx self.weights = SinusoidalPositionalEmbedding.get_embedding( init_size, embedding_dim, padding_idx, ) self.register_buffer('_float_tensor', torch.FloatTensor(1)) @staticmethod def get_embedding(num_embeddings, embedding_dim, padding_idx=None): """Build sinusoidal embeddings. This matches the implementation in tensor2tensor, but differs slightly from the description in Section 3.5 of "Attention Is All You Need". """ half_dim = embedding_dim // 2 emb = math.log(10000) / (half_dim - 1) emb = torch.exp(torch.arange(half_dim, dtype=torch.float) * -emb) emb = torch.arange(num_embeddings, dtype=torch.float).unsqueeze(1) * emb.unsqueeze(0) emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=1).view(num_embeddings, -1) if embedding_dim % 2 == 1: # zero pad emb = torch.cat([emb, torch.zeros(num_embeddings, 1)], dim=1) if padding_idx is not None: emb[padding_idx, :] = 0 return emb def forward(self, input, incremental_state=None, timestep=None, positions=None, **kwargs): """Input is expected to be of size [bsz x seqlen].""" bsz, seq_len = input.shape[:2] max_pos = self.padding_idx + 1 + seq_len if self.weights is None or max_pos > self.weights.size(0): # recompute/expand embeddings if needed self.weights = SinusoidalPositionalEmbedding.get_embedding( max_pos, self.embedding_dim, self.padding_idx, ) self.weights = self.weights.to(self._float_tensor) if incremental_state is not None: # positions is the same for every token when decoding a single step pos = timestep.view(-1)[0] + 1 if timestep is not None else seq_len return self.weights[self.padding_idx + pos, :].expand(bsz, 1, -1) positions = make_positions(input, self.padding_idx) if positions is None else positions return self.weights.index_select(0, positions.view(-1)).view(bsz, seq_len, -1).detach() def max_positions(self): """Maximum number of supported positions.""" return int(1e5) # an arbitrary large number class RotaryEmbeddings(nn.Module): cos: torch.Tensor sin: torch.Tensor theta: torch.Tensor def __init__( self, width: int, *, seq_len: int = 4000, base: int = 10000, device: Optional[torch.device] = None, ): """Rotary embeddings (Su et al., 2021) layer. The rotary embedding will be precomputed for up to 'seq _len' positions. The embedding will be recomputed when a longer sequence is found in the input. :param width: Rotary embedding dimensionality, must be even. :param seq_len: Number of positons to initially precompute. :param base: The base used for Θ_i, determines the cycle length of the embeddings. :param device: Device on which the module is to be initialized. """ super().__init__() if width % 2: raise ValueError(f"Width of rotary embeddings must be even, was: {width}") # Ignore allocations on the meta device as we don't persist our buffer, # i.e., we don't expect the backing tensor to be replaced with pretrained weights. if device is not None and device.type == "meta": device = None # Θ_i = 10000^(-2(i-1)/d) theta = torch.pow( base, -torch.arange(0, width, 2, dtype=torch.float, device=device) / width ) self.register_buffer("theta", theta, persistent=False) self._create_rotary_embed(width=width, length=seq_len) def _create_rotary_embed(self, *, width: int, length: int): # mΘ position = torch.arange(length, device=self.theta.device).unsqueeze(1) m_theta = position * self.theta.unsqueeze(0) # We apply both sin and cos twice (see Eq 15, 34), but the ordering # is changed for compatibility with most common implementations. m_theta = torch.cat([m_theta, m_theta], dim=-1) re_cos = m_theta.cos().view([length, width]) re_sin = m_theta.sin().view([length, width]) self.register_buffer("cos", re_cos, persistent=False) self.register_buffer("sin", re_sin, persistent=False) def _rotate(self, input: torch.Tensor): """Rotate the input tensor by half of its innermost width. input (Tensor): array to rotate. RETURNS (Tensor): rotated array. Shapes: input - (..., width) output - (..., width) """ half_idx = input.shape[-1] // 2 input_1 = -input[..., half_idx:] input_2 = input[..., :half_idx] return torch.cat([input_1, input_2], dim=-1) def forward(self, input: torch.Tensor, *, positions: Optional[torch.Tensor] = None): """ Apply rotary embeddings to an array. :param input: Array to apply the rotary embeddings to. :param positions: positions of the inputs. If no positions are provided, they are assumed to be [0, seq_len). :return: Array with the rotary embeddings applied. Shapes: input - (batch_size, num_heads, seq_len, width_per_head) positions - (batch_size, seq_len) output - (batch_size, num_heads, seq_len, width_per_head) """ batch_size, _, seq_len, width = input.shape if positions is None: # Fastpath: positions from [0..seq_len), avoid indexing. if self.cos.size(-2) < seq_len: self._create_rotary_embed(width=width, length=seq_len) rot_cos = self.cos[:seq_len, :].view(1, 1, seq_len, width) rot_sin = self.sin[:seq_len, :].view(1, 1, seq_len, width) else: max_len = int(positions.max()) + 1 if self.cos.size(-2) < max_len: self._create_rotary_embed(width=width, length=max_len) # Flatten positions to index cos/sin arrays, then unflatten. # # Example shapes: # # positions_flat - (batch_size * seq_len) # self.cos - (max_len, width) # rot_cos - (batch_size, seq_len, width) positions_flat = positions.view(-1) rot_cos = self.cos[positions_flat].view(batch_size, 1, seq_len, width) rot_sin = self.sin[positions_flat].view(batch_size, 1, seq_len, width) # Eq 34 with ordering changed for compatibility. return rot_cos * input + rot_sin * self._rotate(input) class RotMultiheadAttention(MultiheadAttention): def __init__(self, embed_dim, num_heads, kdim=None, vdim=None, dropout=0., bias=True, add_bias_kv=False, add_zero_attn=False, self_attention=False, encoder_decoder_attention=False): super().__init__(embed_dim, num_heads, kdim=kdim, vdim=vdim, dropout=dropout, bias=bias, add_bias_kv=add_bias_kv, add_zero_attn=add_zero_attn, self_attention=self_attention, encoder_decoder_attention=encoder_decoder_attention) self.rotary_embeds = RotaryEmbeddings(width=embed_dim // num_heads) def forward( self, query, key, value, spk_pos_ids_flat=None, key_padding_mask=None, incremental_state=None, need_weights=True, static_kv=False, attn_mask=None, before_softmax=False, need_head_weights=False, enc_dec_attn_constraint_mask=None, reset_attn_weight=None ): """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 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 '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: # self-attention q, k, v = self.in_proj_qkv(query) elif self.encoder_decoder_attention: # encoder-decoder attention q = self.in_proj_q(query) if key is None: assert value is None k = v = None else: k = self.in_proj_k(key) v = self.in_proj_v(key) else: q = self.in_proj_q(query) k = self.in_proj_k(key) v = self.in_proj_v(value) q = 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) # Apply rot embedding and store incremental_state q = self.rotary_embeds(q[None, :], positions=spk_pos_ids_flat)[0] 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'].view(bsz * self.num_heads, -1, self.head_dim) if static_kv: k = prev_key else: k = torch.cat((prev_key, k), dim=1) if 'prev_value' in saved_state: prev_value = saved_state['prev_value'].view(bsz * self.num_heads, -1, self.head_dim) if static_kv: v = prev_value else: v = torch.cat((prev_value, v), dim=1) saved_state['prev_key'], saved_state['prev_value'] = k.view(bsz, self.num_heads, -1, self.head_dim), v.view( bsz, self.num_heads, -1, self.head_dim) self._set_input_buffer(incremental_state, saved_state) if incremental_state is not None: key_pos = torch.arange(k.shape[-2], device=q.device).unsqueeze(0) else: key_pos = spk_pos_ids_flat k = self.rotary_embeds(k[None, :], positions=key_pos)[0] 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.shape == torch.Size([]): 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: 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 attn_mask is not None: if len(attn_mask.shape) == 2: attn_mask = attn_mask.unsqueeze(0) elif len(attn_mask.shape) == 3: attn_mask = attn_mask[:, None].repeat([1, self.num_heads, 1, 1]).reshape( bsz * self.num_heads, tgt_len, src_len) attn_weights = attn_weights + attn_mask if enc_dec_attn_constraint_mask is not None: # bs x head x L_kv attn_weights = attn_weights.view(bsz, self.num_heads, tgt_len, src_len) attn_weights = attn_weights.masked_fill( enc_dec_attn_constraint_mask.unsqueeze(2).bool(), -1e8, ) attn_weights = attn_weights.view(bsz * self.num_heads, tgt_len, src_len) 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) attn_weights = attn_weights.masked_fill( key_padding_mask.unsqueeze(1).unsqueeze(2), -1e8, ) attn_weights = attn_weights.view(bsz * self.num_heads, tgt_len, src_len) attn_logits = attn_weights.view(bsz, self.num_heads, tgt_len, src_len) if before_softmax: return attn_weights, v attn_weights_float = softmax(attn_weights, dim=-1) attn_weights = attn_weights_float.type_as(attn_weights) attn_probs = F.dropout(attn_weights_float.type_as(attn_weights), p=self.dropout, training=self.training) if reset_attn_weight is not None: if reset_attn_weight: self.last_attn_probs = attn_probs.detach() else: assert self.last_attn_probs is not None attn_probs = self.last_attn_probs attn = torch.bmm(attn_probs, v) assert list(attn.size()) == [bsz * self.num_heads, tgt_len, self.head_dim] attn = attn.transpose(0, 1).contiguous().view(tgt_len, bsz, embed_dim) attn = self.out_proj(attn) 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) else: attn_weights = None return attn, (attn_weights, attn_logits) class RotMultiheadAttention2(MultiheadAttention): def __init__(self, embed_dim, num_heads, kdim=None, vdim=None, dropout=0., bias=True, add_bias_kv=False, add_zero_attn=False, self_attention=False, encoder_decoder_attention=False): super().__init__(embed_dim, num_heads, kdim=kdim, vdim=vdim, dropout=dropout, bias=bias, add_bias_kv=add_bias_kv, add_zero_attn=add_zero_attn, self_attention=self_attention, encoder_decoder_attention=encoder_decoder_attention) self.rotary_embeds = RotaryEmbeddings(width=embed_dim // num_heads) def forward( self, query, key, value, spk_pos_ids_flat=None, key_padding_mask=None, incremental_state=None, need_weights=True, static_kv=False, attn_mask=None, before_softmax=False, need_head_weights=False, enc_dec_attn_constraint_mask=None, reset_attn_weight=None ): """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 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 '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: # self-attention q, k, v = self.in_proj_qkv(query) elif self.encoder_decoder_attention: # encoder-decoder attention q = self.in_proj_q(query) if key is None: assert value is None k = v = None else: k = self.in_proj_k(key) v = self.in_proj_v(key) else: q = self.in_proj_q(query) k = self.in_proj_k(key) v = self.in_proj_v(value) 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) # Apply rot embedding and store incremental_state q = self.rotary_embeds(q[None, :], positions=spk_pos_ids_flat)[0] 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'].view(bsz * self.num_heads, -1, self.head_dim) if static_kv: k = prev_key else: k = torch.cat((prev_key, k), dim=1) if 'prev_value' in saved_state: prev_value = saved_state['prev_value'].view(bsz * self.num_heads, -1, self.head_dim) if static_kv: v = prev_value else: v = torch.cat((prev_value, v), dim=1) saved_state['prev_key'], saved_state['prev_value'] = k.view(bsz, self.num_heads, -1, self.head_dim), v.view( bsz, self.num_heads, -1, self.head_dim) self._set_input_buffer(incremental_state, saved_state) key_pos = torch.arange(k.shape[-2], device=q.device).unsqueeze(0) k = self.rotary_embeds(k[None, :], positions=key_pos)[0] 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.shape == torch.Size([]): 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 attn_mask is not None: if len(attn_mask.shape) == 2: attn_mask = attn_mask.unsqueeze(0) elif len(attn_mask.shape) == 3: attn_mask = attn_mask[:, None].repeat([1, self.num_heads, 1, 1]).reshape( bsz * self.num_heads, tgt_len, src_len) attn = torch.nn.functional.scaled_dot_product_attention( q, k, v, attn_mask=attn_mask, dropout_p=0, is_causal=False) assert list(attn.size()) == [bsz * self.num_heads, tgt_len, self.head_dim] attn = attn.transpose(0, 1).contiguous().view(tgt_len, bsz, embed_dim) attn_logits = None attn_weights = None return attn, (attn_weights, attn_logits) class RotDecSALayer(nn.Module): def __init__(self, c, num_heads, dropout, attention_dropout=0.1, relu_dropout=0.1, kernel_size=9, ffn_hidden_size=1024, act='gelu', post_ln=False): super().__init__() self.c = c self.dropout = dropout self.layer_norm1 = LayerNorm(c) self.self_attn = RotMultiheadAttention( c, num_heads, self_attention=True, dropout=attention_dropout, bias=False ) self.layer_norm2 = LayerNorm(c) self.ffn = TransformerFFNLayer( c, ffn_hidden_size, padding='LEFT', kernel_size=kernel_size, dropout=relu_dropout, act=act) self.post_ln = post_ln def forward( self, x, encoder_out=None, encoder_padding_mask=None, incremental_state=None, self_attn_mask=None, self_attn_padding_mask=None, attn_out=None, reset_attn_weight=None, spk_pos_ids_flat=None, **kwargs, ): layer_norm_training = kwargs.get('layer_norm_training', None) if layer_norm_training is not None: self.layer_norm1.training = layer_norm_training self.layer_norm2.training = layer_norm_training residual = x if not self.post_ln: x = self.layer_norm1(x) x, (attn_weights, _) = self.self_attn( query=x, key=x, value=x, key_padding_mask=self_attn_padding_mask, incremental_state=incremental_state, attn_mask=self_attn_mask, spk_pos_ids_flat=spk_pos_ids_flat ) x = F.dropout(x, self.dropout, training=self.training) x = residual + x if self.post_ln: x = self.layer_norm1(x) residual = x if not self.post_ln: x = self.layer_norm2(x) x = self.ffn(x, incremental_state=incremental_state) x = F.dropout(x, self.dropout, training=self.training) x = residual + x if self.post_ln: x = self.layer_norm2(x) return x, attn_weights def clear_buffer(self, input, encoder_out=None, encoder_padding_mask=None, incremental_state=None): self.encoder_attn.clear_buffer(incremental_state) self.ffn.clear_buffer(incremental_state) def set_buffer(self, name, tensor, incremental_state): return set_incremental_state(self, incremental_state, name, tensor) class RotDecSALayer2(RotDecSALayer): def __init__(self, c, num_heads, dropout, attention_dropout=0.1, relu_dropout=0.1, kernel_size=9, ffn_hidden_size=1024, act='gelu', post_ln=False): super().__init__(c, num_heads, dropout, attention_dropout, relu_dropout, kernel_size, ffn_hidden_size, act, post_ln) self.self_attn = RotMultiheadAttention2( c, num_heads, self_attention=True, dropout=attention_dropout, bias=False ) class RotTransformerDecoderLayer(nn.Module): def __init__(self, hidden_size, dropout, kernel_size=9, num_heads=8, ffn_hidden_size=1024, post_ln=False, op_version=1): super().__init__() self.hidden_size = hidden_size self.dropout = dropout self.num_heads = num_heads if op_version == 1: self.op = RotDecSALayer( hidden_size, num_heads, dropout=dropout, attention_dropout=0.0, relu_dropout=dropout, kernel_size=kernel_size, ffn_hidden_size=ffn_hidden_size, post_ln=post_ln) else: self.op = RotDecSALayer2( hidden_size, num_heads, dropout=dropout, attention_dropout=0.0, relu_dropout=dropout, kernel_size=kernel_size, ffn_hidden_size=ffn_hidden_size, post_ln=post_ln) def forward(self, x, **kwargs): return self.op(x, **kwargs) def clear_buffer(self, *args): return self.op.clear_buffer(*args) def set_buffer(self, *args): return self.op.set_buffer(*args)