torchscale/torchscale/architecture/encoder.py

390 lines
13 KiB
Python

# Copyright (c) 2022 Microsoft
# Licensed under The MIT License [see LICENSE for details]
import math
import numpy as np
import torch
import torch.nn as nn
from fairscale.nn import checkpoint_wrapper, wrap
try:
from apex.normalization import FusedLayerNorm as LayerNorm
except ModuleNotFoundError:
from torch.nn import LayerNorm
from torchscale.architecture.utils import init_bert_params
from torchscale.component.droppath import DropPath
from torchscale.component.feedforward_network import FeedForwardNetwork, make_experts
from torchscale.component.multihead_attention import MultiheadAttention
from torchscale.component.multiway_network import MultiwayWrapper, set_split_position
from torchscale.component.relative_position_bias import RelativePositionBias
from torchscale.component.xmoe.moe_layer import MOELayer
from torchscale.component.xmoe.routing import Top1Gate, Top2Gate
class EncoderLayer(nn.Module):
def __init__(self, args, depth, is_moe_layer=False, is_encoder_decoder=False):
super().__init__()
self.args = args
self.embed_dim = args.encoder_embed_dim
self.self_attn = self.build_self_attention(self.embed_dim, args)
self.self_attn_layer_norm = MultiwayWrapper(args, LayerNorm(self.embed_dim, eps=args.layernorm_eps))
self.dropout_module = torch.nn.Dropout(args.dropout)
if args.drop_path_rate > 0:
drop_path_prob = np.linspace(0, args.drop_path_rate, args.encoder_layers)[
depth
]
self.drop_path = DropPath(drop_path_prob)
else:
self.drop_path = None
self.normalize_before = args.encoder_normalize_before
self.is_moe_layer = is_moe_layer
self.ffn_dim = args.encoder_ffn_embed_dim
if not self.is_moe_layer:
self.ffn = MultiwayWrapper(
args,
self.build_ffn(
self.embed_dim,
self.args,
),
)
else:
assert not self.args.multiway
if args.moe_top1_expert:
gate = Top1Gate(
self.embed_dim,
args.moe_expert_count,
use_fp32=args.moe_gating_use_fp32,
moe_eval_capacity_token_fraction=args.moe_eval_capacity_token_fraction,
use_xmoe=args.use_xmoe,
)
else:
gate = Top2Gate(
self.embed_dim,
args.moe_expert_count,
args.moe_gating_use_fp32,
args.moe_second_expert_policy,
args.moe_normalize_gate_prob_before_dropping,
args.moe_eval_capacity_token_fraction,
use_xmoe=args.use_xmoe,
)
experts = make_experts(args, self.embed_dim, self.ffn_dim)
self.moe_layer = MOELayer(gate, experts, args)
self.final_layer_norm = MultiwayWrapper(args, LayerNorm(self.embed_dim, eps=args.layernorm_eps))
if args.deepnorm:
if is_encoder_decoder:
self.alpha = (
math.pow(
math.pow(args.encoder_layers, 4) * args.decoder_layers, 0.0625
)
* 0.81
)
else:
self.alpha = math.pow(2.0 * args.encoder_layers, 0.25)
else:
self.alpha = 1.0
def build_ffn(self, embed_dim, args):
return FeedForwardNetwork(
embed_dim,
self.ffn_dim,
args.activation_fn,
args.dropout,
args.activation_dropout,
args.layernorm_eps,
args.subln,
)
def build_self_attention(self, embed_dim, args):
return MultiheadAttention(
args,
embed_dim,
args.encoder_attention_heads,
dropout=args.attention_dropout,
self_attention=True,
encoder_decoder_attention=False,
subln=args.subln,
)
def residual_connection(self, x, residual):
return residual * self.alpha + x
def forward(self, x, encoder_padding_mask, attn_mask=None, rel_pos=None, multiway_split_position=None, incremental_state=None):
if multiway_split_position is not None:
assert self.args.multiway
self.apply(set_split_position(multiway_split_position))
if attn_mask is not None:
attn_mask = attn_mask.masked_fill(attn_mask.to(torch.bool), -1e8)
residual = x
if self.normalize_before:
x = self.self_attn_layer_norm(x)
x, _ = self.self_attn(
query=x,
key=x,
value=x,
key_padding_mask=encoder_padding_mask,
attn_mask=attn_mask,
rel_pos=rel_pos,
incremental_state=incremental_state,
)
x = self.dropout_module(x)
if self.drop_path is not None:
x = self.drop_path(x)
x = self.residual_connection(x, residual)
if not self.normalize_before:
x = self.self_attn_layer_norm(x)
residual = x
if self.normalize_before:
x = self.final_layer_norm(x)
if not self.is_moe_layer:
x = self.ffn(x)
l_aux = None
else:
x = x.transpose(0, 1)
x, l_aux = self.moe_layer(x)
x = x.transpose(0, 1)
if self.drop_path is not None:
x = self.drop_path(x)
x = self.residual_connection(x, residual)
if not self.normalize_before:
x = self.final_layer_norm(x)
return x, l_aux
class Encoder(nn.Module):
def __init__(
self,
args,
embed_tokens=None,
embed_positions=None,
output_projection=None,
is_encoder_decoder=False,
**kwargs
):
self.args = args
super().__init__(**kwargs)
self.dropout_module = torch.nn.Dropout(args.dropout)
embed_dim = args.encoder_embed_dim
self.embed_scale = 1.0 if args.no_scale_embedding else math.sqrt(embed_dim)
self.embed_tokens = embed_tokens
self.embed_positions = embed_positions
if (
output_projection is None
and not is_encoder_decoder
and not args.no_output_layer
and args.vocab_size > 0
):
self.output_projection = self.build_output_projection(args)
else:
self.output_projection = output_projection
if args.layernorm_embedding:
self.layernorm_embedding = MultiwayWrapper(
args, LayerNorm(embed_dim, eps=args.layernorm_eps), dim=1
)
else:
self.layernorm_embedding = None
self.layers = nn.ModuleList([])
moe_freq = args.moe_freq
for i in range(args.encoder_layers):
is_moe_layer = moe_freq != 0 and (i + 1) % moe_freq == 0
self.layers.append(
self.build_encoder_layer(
args,
depth=i,
is_moe_layer=is_moe_layer,
is_encoder_decoder=is_encoder_decoder,
)
)
self.num_layers = len(self.layers)
if args.encoder_normalize_before and args.normalize_output:
self.layer_norm = MultiwayWrapper(args, LayerNorm(embed_dim, eps=args.layernorm_eps))
else:
self.layer_norm = None
if args.rel_pos_buckets > 0 and args.max_rel_pos > 0:
self.relative_position = RelativePositionBias(
num_buckets=args.rel_pos_buckets,
max_distance=args.max_rel_pos,
n_heads=args.encoder_attention_heads,
)
else:
self.relative_position = None
if args.bert_init:
self.apply(init_bert_params)
if args.deepnorm:
if is_encoder_decoder:
init_scale = (
math.pow(
math.pow(args.encoder_layers, 4) * args.decoder_layers, 0.0625
)
/ 1.15
)
else:
init_scale = math.pow(8.0 * args.encoder_layers, 0.25)
for name, p in self.named_parameters():
if (
"fc1" in name
or "fc2" in name
or "out_proj" in name
or "v_proj" in name
):
p.data.div_(init_scale)
if args.subln:
if is_encoder_decoder:
init_scale = math.sqrt(
math.log(3 * args.decoder_layers)
* math.log(2 * args.encoder_layers)
/ 3
)
else:
init_scale = math.sqrt(math.log(args.encoder_layers * 2))
for name, p in self.named_parameters():
if (
"fc1" in name
or "fc2" in name
or "out_proj" in name
or "v_proj" in name
):
p.data.mul_(init_scale)
def build_output_projection(
self,
args,
):
if args.share_encoder_input_output_embed:
assert args.encoder_embedding_type == "language"
output_projection = torch.nn.Linear(
self.embed_tokens.weight.shape[1],
self.embed_tokens.weight.shape[0],
bias=False,
)
output_projection.weight = self.embed_tokens.weight
else:
output_projection = torch.nn.Linear(
args.encoder_embed_dim, args.vocab_size, bias=False
)
torch.nn.init.normal_(
output_projection.weight, mean=0, std=args.encoder_embed_dim**-0.5
)
return output_projection
def build_encoder_layer(
self, args, depth, is_moe_layer=False, is_encoder_decoder=False
):
layer = EncoderLayer(
args,
depth,
is_moe_layer=is_moe_layer,
is_encoder_decoder=is_encoder_decoder,
)
if args.checkpoint_activations:
layer = checkpoint_wrapper(layer)
if args.fsdp:
layer = wrap(layer)
return layer
def forward_embedding(
self,
src_tokens,
token_embedding=None,
positions=None,
):
if token_embedding is None:
token_embedding = self.embed_tokens(src_tokens)
x = embed = self.embed_scale * token_embedding
if self.embed_positions is not None:
if src_tokens is not None:
x = embed + self.embed_positions(src_tokens, positions=positions)
else:
x = embed + self.embed_positions(x, positions=positions)
if self.layernorm_embedding is not None:
x = self.layernorm_embedding(x)
x = self.dropout_module(x)
return x, embed
def forward(
self,
src_tokens,
encoder_padding_mask=None,
attn_mask=None,
return_all_hiddens=False,
token_embeddings=None,
multiway_split_position=None,
features_only=False,
incremental_state=None,
positions=None,
**kwargs
):
assert src_tokens is not None or token_embeddings is not None
if multiway_split_position is not None:
assert self.args.multiway
self.apply(set_split_position(multiway_split_position))
x, encoder_embedding = self.forward_embedding(src_tokens, token_embeddings, positions)
if encoder_padding_mask is not None:
x = x * (1 - encoder_padding_mask.unsqueeze(-1).type_as(x))
encoder_states = []
if return_all_hiddens:
encoder_states.append(x)
rel_pos_bias = None
if self.relative_position is not None:
rel_pos_bias = self.relative_position(
batch_size=x.size(0), qlen=x.size(1), klen=x.size(1)
)
# incremental_state is not None during inference if we use the bidirectional encoder as a generator as in s2s-ft (https://arxiv.org/abs/2110.13640)
l_aux = []
for idx, layer in enumerate(self.layers):
x, l_aux_i = layer(
x,
encoder_padding_mask=encoder_padding_mask if incremental_state is None else None,
attn_mask=attn_mask,
rel_pos=rel_pos_bias,
multiway_split_position=multiway_split_position,
incremental_state=incremental_state[idx] if incremental_state is not None else None,
)
if return_all_hiddens:
assert encoder_states is not None
encoder_states.append(x)
l_aux.append(l_aux_i)
if self.layer_norm is not None:
x = self.layer_norm(x)
if not features_only and self.output_projection is not None:
x = self.output_projection(x)
return {
"encoder_out": x,
"encoder_embedding": encoder_embedding,
"encoder_padding_mask": encoder_padding_mask,
"encoder_states": encoder_states,
"l_aux": l_aux,
}