sinkhorn transformer
0.11.4
这是稀疏 Sinkhorn Attention 中概述的工作的复制品，并进行了额外的增强。
它包括一个参数化排序网络，使用 Sinkhorn 标准化对排列矩阵进行采样，该排列矩阵将最相关的键桶与查询桶相匹配。
这项工作还引入了可逆网络和前馈分块（从 Reformer 引入的概念），以进一步节省内存。
204k 代币（演示目的）
$ pip install sinkhorn_transformer基于 Sinkhorn Transformer 的语言模型
import torch
from sinkhorn_transformer import SinkhornTransformerLM
model = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 1024 ,
heads = 8 ,
depth = 12 ,
max_seq_len = 8192 ,
bucket_size = 128 , # size of the buckets
causal = False , # auto-regressive or not
n_sortcut = 2 , # use sortcut to reduce memory complexity to linear
n_top_buckets = 2 , # sort specified number of key/value buckets to one query bucket. paper is at 1, defaults to 2
ff_chunks = 10 , # feedforward chunking, from Reformer paper
reversible = True , # make network reversible, from Reformer paper
emb_dropout = 0.1 , # embedding dropout
ff_dropout = 0.1 , # feedforward dropout
attn_dropout = 0.1 , # post attention dropout
attn_layer_dropout = 0.1 , # post attention layer dropout
layer_dropout = 0.1 , # add layer dropout, from 'Reducing Transformer Depth on Demand' paper
weight_tie = True , # tie layer parameters, from Albert paper
emb_dim = 128 , # embedding factorization, from Albert paper
dim_head = 64 , # be able to fix the dimension of each head, making it independent of the embedding dimension and the number of heads
ff_glu = True , # use GLU in feedforward, from paper 'GLU Variants Improve Transformer'
n_local_attn_heads = 2 , # replace N heads with local attention, suggested to work well from Routing Transformer paper
pkm_layers = ( 4 , 7 ), # specify layers to use product key memory. paper shows 1 or 2 modules near the middle of the transformer is best
pkm_num_keys = 128 , # defaults to 128, but can be increased to 256 or 512 as memory allows
)
x = torch . randint ( 0 , 20000 , ( 1 , 2048 ))
model ( x ) # (1, 2048, 20000)一个普通的Sinkhorn Transformer，层层的Sinkhorn注意力
import torch
from sinkhorn_transformer import SinkhornTransformer
model = SinkhornTransformer (
dim = 1024 ,
heads = 8 ,
depth = 12 ,
bucket_size = 128
)
x = torch . randn ( 1 , 2048 , 1024 )
model ( x ) # (1, 2048, 1024)Sinkhorn 编码器/解码器变压器
import torch
from sinkhorn_transformer import SinkhornTransformerLM
DE_SEQ_LEN = 4096
EN_SEQ_LEN = 4096
enc = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 512 ,
depth = 6 ,
heads = 8 ,
bucket_size = 128 ,
max_seq_len = DE_SEQ_LEN ,
reversible = True ,
return_embeddings = True
). cuda ()
dec = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 512 ,
depth = 6 ,
causal = True ,
bucket_size = 128 ,
max_seq_len = EN_SEQ_LEN ,
receives_context = True ,
context_bucket_size = 128 , # context key / values can be bucketed differently
reversible = True
). cuda ()
x = torch . randint ( 0 , 20000 , ( 1 , DE_SEQ_LEN )). cuda ()
y = torch . randint ( 0 , 20000 , ( 1 , EN_SEQ_LEN )). cuda ()
x_mask = torch . ones_like ( x ). bool (). cuda ()
y_mask = torch . ones_like ( y ). bool (). cuda ()
context = enc ( x , input_mask = x_mask )
dec ( y , context = context , input_mask = y_mask , context_mask = x_mask ) # (1, 4096, 20000) 默认情况下，如果给定的输入不是桶大小的倍数，模型会抱怨。为了避免每次都进行相同的填充计算，您可以使用帮助器Autopadder类。如果给定的话，它也会为您处理input_mask 。还支持上下文键/值和掩码。
import torch
from sinkhorn_transformer import SinkhornTransformerLM
from sinkhorn_transformer import Autopadder
model = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 1024 ,
heads = 8 ,
depth = 12 ,
max_seq_len = 2048 ,
bucket_size = 128 ,
causal = True
)
model = Autopadder ( model , pad_left = True ) # autopadder will fetch the bucket size and autopad input
x = torch . randint ( 0 , 20000 , ( 1 , 1117 )) # odd sequence length
model ( x ) # (1, 1117, 20000) 该存储库与论文有所不同，现在使用注意力机制代替原来的排序网+gumbel sinkhorn 采样。我还没有发现性能上有明显的差异，新的方案允许我将网络推广到灵活的序列长度。如果您想尝试 Sinkhorn，请使用以下设置，该设置仅适用于非因果网络。
import torch
from sinkhorn_transformer import SinkhornTransformerLM
model = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 1024 ,
heads = 8 ,
depth = 12 ,
bucket_size = 128 ,
max_seq_len = 8192 ,
use_simple_sort_net = True , # turn off attention sort net
sinkhorn_iter = 7 , # number of sinkhorn iterations - default is set at reported best in paper
n_sortcut = 2 , # use sortcut to reduce complexity to linear time
temperature = 0.75 , # gumbel temperature - default is set at reported best in paper
non_permutative = False , # allow buckets of keys to be sorted to queries more than once
)
x = torch . randint ( 0 , 20000 , ( 1 , 8192 ))
model ( x ) # (1, 8192, 20000) 要看到使用 PKM 的好处，必须将值的学习率设置为高于其余参数。 （建议为1e-2 ）
您可以按照此处的说明正确设置 https://github.com/lucidrains/product-key-memory#learning-rates
Sinkhorn 在固定长度序列上进行训练时，似乎无法从头开始解码序列，这主要是由于当桶部分填充填充令牌时排序网络难以泛化。
幸运的是，我想我已经找到了一个简单的解决方案。在训练过程中，对于因果网络，随机截断序列并强制排序网络进行泛化。我为AutoregressiveWrapper实例提供了一个标志（ randomly_truncate_sequence ），以简化此操作。
import torch
from sinkhorn_transformer import SinkhornTransformerLM , AutoregressiveWrapper
model = SinkhornTransformerLM (
num_tokens = 20000 ,
dim = 1024 ,
heads = 8 ,
depth = 12 ,
bucket_size = 75 ,
max_seq_len = 8192 ,
causal = True
)
model = AutoregressiveWrapper ( model )
x = torch . randint ( 0 , 20000 , ( 1 , 8192 ))
loss = model ( x , return_loss = True , randomly_truncate_sequence = True ) # (1, 8192, 20000)如果有人找到更好的解决方案，我愿意接受建议。
因果排序网络存在一个潜在的问题，其中将过去的哪些键/值存储桶排序到存储桶的决定仅取决于第一个令牌，而不依赖于其余令牌（由于分桶方案并防止未来的泄漏）过去的）。
我试图通过将一半的头向左旋转桶大小 - 1 来缓解这个问题，从而将最后一个令牌提升为第一个。这也是AutoregressiveWrapper在训练期间默认使用左填充的原因，以始终确保序列中的最后一个标记对要检索的内容有发言权。
如果有人找到了更干净的解决方案，请在问题中告诉我。
@misc { tay2020sparse ,
title = { Sparse Sinkhorn Attention } ,
author = { Yi Tay and Dara Bahri and Liu Yang and Donald Metzler and Da-Cheng Juan } ,
year = { 2020 } ,
url. = { https://arxiv.org/abs/2002.11296 }
} @inproceedings { kitaev2020reformer ,
title = { Reformer: The Efficient Transformer } ,
author = { Nikita Kitaev and Lukasz Kaiser and Anselm Levskaya } ,
booktitle = { International Conference on Learning Representations } ,
year = { 2020 } ,
url = { https://openreview.net/forum?id=rkgNKkHtvB }
} @misc { lan2019albert ,
title = { ALBERT: A Lite BERT for Self-supervised Learning of Language Representations } ,
author = { Zhenzhong Lan and Mingda Chen and Sebastian Goodman and Kevin Gimpel and Piyush Sharma and Radu Soricut } ,
year = { 2019 } ,
url = { https://arxiv.org/abs/1909.11942 }
} @misc { shazeer2020glu ,
title = { GLU Variants Improve Transformer } ,
author = { Noam Shazeer } ,
year = { 2020 } ,
url = { https://arxiv.org/abs/2002.05202 }
} @misc { roy*2020efficient ,
title = { Efficient Content-Based Sparse Attention with Routing Transformers } ,
author = { Aurko Roy* and Mohammad Taghi Saffar* and David Grangier and Ashish Vaswani } ,
year = { 2020 } ,
url = { https://openreview.net/forum?id=B1gjs6EtDr }
} @inproceedings { fan2020reducing ,
title = { Reducing Transformer Depth on Demand with Structured Dropout } ,
author = { Angela Fan and Edouard Grave and Armand Joulin } ,
booktitle = { International Conference on Learning Representations } ,
year = { 2020 } ,
url = { https://openreview.net/forum?id=SylO2yStDr }
} @misc { lample2019large ,
title = { Large Memory Layers with Product Keys } ,
author = { Guillaume Lample and Alexandre Sablayrolles and Marc'Aurelio Ranzato and Ludovic Denoyer and Hervé Jégou } ,
year = { 2019 } ,
eprint = { 1907.05242 } ,
archivePrefix = { arXiv }
} @misc { bhojanapalli2020lowrank ,
title = { Low-Rank Bottleneck in Multi-head Attention Models } ,
author = { Srinadh Bhojanapalli and Chulhee Yun and Ankit Singh Rawat and Sashank J. Reddi and Sanjiv Kumar } ,
year = { 2020 } ,
eprint = { 2002.07028 }
}
