lucidrains reformer-pytorch issues

lucidrains / reformer-pytorch

Reformer, the efficient Transformer, in Pytorch

MIT License

2.13k stars 256 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Is the unsort logits necessary?

#156 Croooooow opened 6 months ago
0
how to get the dataset of enwik8.gz ?

#155 SeekPoint opened 11 months ago
0
ImportError: cannot import name 'amp' from 'apex'

#154 amrzv opened 1 year ago
1
not find axial_positional_embedding.py

#153 wuyulong666 closed 1 year ago
2
Bump torch from 1.4.0 to 1.13.1 in /pretraining

#152 dependabot[bot] opened 1 year ago
0
Bump transformers from 2.4.1 to 4.30.0 in /pretraining

#151 dependabot[bot] closed 1 year ago
0
Bump numpy from 1.18.1 to 1.22.0 in /pretraining

#150 dependabot[bot] closed 2 years ago
0
FLOPs calculation for LSHSelfAttention in LSH mode and Full attention mode

#149 zaidilyas89 opened 2 years ago
0
half precision on long sequence doesn't work

#148 sklin93 closed 3 years ago
3
error in eval loss & ppl

#147 dar-tau opened 3 years ago
0
Default embeddings are rotary Embeddings not Axialpositional Embeddings as stated by the documentation

#146 samuelgoodall closed 3 years ago
1
About sequence length

#145 czydfj opened 3 years ago
0
fp16 support

#144 1140310118 closed 3 years ago
0
self-supervised.py Training Script is it worthwhile pretokenizing ?

#143 samuelgoodall closed 3 years ago
0
Reversible layers increase memory usage

#142 serkansulun opened 3 years ago
1
Is the argument max_seq_len in Reformer() used?

#141 e4exp closed 3 years ago
1
Error while testing Code in Readme

#140 samuelgoodall closed 3 years ago
2
Class token implementation

#139 karttikeya opened 3 years ago
1
why lucidrains use reformerlm as pre train model?

#138 jiangyums opened 3 years ago
0
Image as input

#137 fedeloper opened 3 years ago
0
The exact English pretraining data and Chinese pretraining data that are exact same to the BERT paper's pretraining data.

#136 guotong1988 opened 3 years ago
0
Music Reformer results (some problems)

#135 asigalov61 opened 3 years ago
1
TOKEN_SELF_ATTN_VALUE and QK attention

#134 lcmeng opened 3 years ago
0
Are `emb_dim` and `dim` equal by default

#133 rookiez7 closed 3 years ago
1
about input_mask

#132 rookiez7 closed 3 years ago
0
Consider creating a transformers library

#131 jordiae opened 3 years ago
0
about sequence mask in classification

#130 rookiez7 closed 3 years ago
0
Bug in layer norm

#129 maclandrol closed 3 years ago
1
why not just use tensorflow keras？

#128 lomessa closed 3 years ago
0
how to do sequence classfication

#127 rookiez7 closed 3 years ago
0
How to evaluate the BLEU score and train on other datasets?

#126 juntang-zhuang opened 3 years ago
0
Enc-Dec not learning

#125 MemduhG closed 3 years ago
1
Scaling by sqrt(dim) when using tied embeddings

#124 gulnazaki closed 3 years ago
4
Request for help for LSHSelfAttention()

#123 andreabac3 closed 3 years ago
22
Issue about Reformer Encoder Decoder Architecture

#122 Henrykwokkk closed 4 years ago
3
handle the overlap problem in hash vector without rehash

#121 wangyu1997 closed 4 years ago
0
reformer hash_vectors without rehash fixed

#120 wangyu1997 closed 4 years ago
1
reformer hash_vectors without rehash fixed

#119 wangyu1997 closed 4 years ago
0
Fix hash_vertor without rehash

#118 wangyu1997 closed 4 years ago
0
what is the meaning of "buckets_and_t = seqlen * buckets + (ticker % seqlen)" ?

#117 onedreame closed 4 years ago
1
Can I use cpu for ReformerEncDec?

#116 youlive789 closed 4 years ago
1
Autopadder not working Reformer

#115 jamarju closed 4 years ago
1
ReformerEncDec generate slow function

#114 allhelllooz closed 4 years ago
4
Some questions about the implement.

#113 WhatAShot closed 4 years ago
2
batch size incompatible for mem_kv

#112 keroro824 closed 4 years ago
1
q and k different?

#111 fishWangY closed 4 years ago
1
Does the framework support fine-tuning with other pre-training models?

#110 solemnrole closed 4 years ago
2
Some questions about class LSHAttention(nn.Module)

#109 L-Hugh closed 4 years ago
4
Feed Forward Dimension for Reformer

#108 anisdismail closed 4 years ago
2
Attn weight size from Recorder

#107 nitinnairk closed 4 years ago
3