PyPI 官网下载 | linear_attention_transformer-0.5.0.tar.gz

## Linear Attention Transformer <img src="./linear-attention.png" width="700px" /> [](https://badge.fury.io/py/linear-attention-transformer) A fully featured Transformer that mixes (QKᵀ)V local attention with Q(KᵀV) global attention (scales linearly with respect to sequence length) for efficient long-range language modeling. ## Install ```bash $ pip install linear-attention-transformer ``` ## Usage Language model ```python import torch from linear_attention_transformer import LinearAttentionTransformerLM model = LinearAttentionTransformerLM( num_tokens = 20000, dim = 512, heads = 8, depth = 1, max_seq_len = 8192, causal = True, # auto-regressive or not blindspot_size = 64, # this gives the q(kv) attention a blindspot of 64 tokens back in the causal case, but gives back an order of magnitude return in memory savings. should be paired with local attention of at least a window size of this setting. setting this to 1 will allow for full q(kv) attention of past n_local_attn_heads = 4, # number of local attention heads for (qk)v attention. this can be a tuple specifying the exact number of local attention heads at that depth local_attn_window_size = 128, # receptive field of the local attention one_kv_head = True, # use one key/value head to save on memory / compute reversible = True, # use reversible nets, from Reformer paper ff_chunks = 2, # feedforward chunking, from Reformer paper psi_fn = lambda x: x.sigmoid() # allows you to modify the psi function used in 'Transformer is RNN' paper ).cuda() x = torch.randint(0, 20000, (1, 8192)).cuda() model(x) # (1, 8192, 512) ``` Transformer ```python import torch from linear_attention_transformer import LinearAttentionTransformer model = LinearAttentionTransformer( dim = 512, heads = 8, depth = 1, max_seq_len = 8192, n_local_attn_heads = 4 ).cuda() x = torch.randn(1, 8192, 512).cuda() model(x) # (1, 8192, 512) ``` Encoder / decoder ```python import torch from linear_attention_transformer import LinearAttentionTransformerLM enc = LinearAttentionTransformerLM( num_tokens = 20000, dim = 512, heads = 8, depth = 6, max_seq_len = 4096, one_kv_head = True, reversible = True, n_local_attn_heads = 4, return_embeddings = True ).cuda() dec = LinearAttentionTransformerLM( num_tokens = 20000, dim = 512, heads = 8, depth = 6, causal = True, max_seq_len = 4096, one_kv_head = True, reversible = True, receives_context = True, n_local_attn_heads = 4 ).cuda() src = torch.randint(0, 20000, (1, 4096)).cuda() src_mask = torch.ones_like(src).bool().cuda() tgt = torch.randint(0, 20000, (1, 4096)).cuda() tgt_mask = torch.ones_like(tgt).bool().cuda() context = enc(src, input_mask = src_mask) logits = dec(tgt, context = context, input_mask = tgt_mask, context_mask = src_mask) ``` ## Images This repository also contains a concise implementation of this efficient attention for images ```python import torch from linear_attention_transformer.images import ImageLinearAttention attn =ImageLinearAttention( chan = 32, heads = 8, key_dim = 64 # can be decreased to 32 for more memory savings ) img = torch.randn(1, 32, 256, 256) attn(img) # (1, 32, 256, 256) ``` ## Citations ```bibtex @inproceedings{katharopoulos-et-al-2020, author = {Katharopoulos, A. and Vyas, A. and Pappas, N. and Fleuret, F.}, title = {Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention}, booktitle = {Proceedings of the International Conference on Machine Learning (ICML)}, year = {2020}, note = {(to appear)} } ``` ```bibtex @article{shen2019efficient, author = {Zhuoran Shen and Mingyuan Zhang and Haiyu Zhao and Shuai Yi and Hongsheng Li}, title = {Efficient Attention: Attention with Linear Complexities}, journal = {CoRR}, volume = {abs/1812.01243}, year = {2018}, url = {http://arxiv.org/abs/1812.01243} } ``` ```bibtex @misc{shazeer2019fast, title = {Fast Transformer Decoding: One Write-Head is All You Need}, author = {Noam Shazeer}, year = {2019}, eprint = {1911.02150}, archivePrefix = {arXiv} } ``` ```bibtex @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} } ```