Issues
- 0
- 0
Why do you pass query, key, and value through the same fc_layer in transformer_xl model?
#151 opened by wonjunchoi-arc - 2
Why pos_seq is in descending order as the input of positional embedding?
#111 opened by GrindstoneLZX - 1
wrong argument order of _update_mems function!
#118 opened by jech2 - 0
About Using
#150 opened by Tuziking - 6
- 0
[W C:\w\b\windows\pytorch\aten\src\ATen\native\cuda\Indexing.cu:963] Warning: masked_fill_ received a mask with dtype torch.uint8, this behavior is now deprecated,please use a mask with dtype torch.bool instead. (function masked_fill__cuda)
#148 opened by Arsmart1 - 0
How to obtain the data?
#147 opened by Arsmart1 - 0
enwiki8 18 layer model .sh file
#145 opened by vasily789 - 1
RelPartialLearnableDecoder vs RelLearnableDecoder
#142 opened by jannessm - 1
- 0
why i-j always>0
#140 opened by scirocc - 1
linux or windows?
#137 opened by lyxwz - 1
Relative Positional Encoding
#138 opened by LarsHill - 0
- 0
运行不起来
#136 opened by lyons-deng - 0
- 1
- 0
Question: why is relative positional encoding computed with length M vs. L+M in the paper ?
#132 opened by gdoras - 0
- 0
Pytorch programs have been killed unexpectedlly
#130 opened by Dinxin - 1
- 0
Difference between ppl and bpc
#128 opened by valofosho - 0
Copyright missing
#127 opened by kazuki-irie - 0
Why use memory with LMShuffledIterator
#126 opened by serkansulun - 1
Sin/Cos concatenation in Positional Embeddings
#119 opened by zainsarwar865 - 3
Different training steps in tf and pytorch
#114 opened by richardbaihe - 0
RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`
#125 opened by demdecuong - 0
Can someone please tell me on what dataset was transformer-XL pre-trained on?
#124 opened by pathak-aman - 1
tf 2.x and python 3.x
#121 opened by kdlin - 0
TF base model memory requirements
#120 opened by tonytan48 - 4
can not reproduce sota wikitext103 results
#112 opened by menghuanlater - 1
Pytorch questions!
#115 opened by garysun1994 - 0
What is the meaning of 'bsz' in mem_transformer.py?
#116 opened by oshindow - 0
fine-tune text classification?
#117 opened by vr25 - 0
Perplexity not changes with tgt_len
#113 opened by bajajahsaas - 2
PositionalEmbedding error
#110 opened by Macielyoung - 2
Bounty: PTB Transformer-xl
#109 opened by srush - 0
论文中的figure1有些看不懂,有大神可以解答一下吗?
#108 opened by heroazhe - 0
- 0
Best settings to train Transformer-XL from scratch
#106 opened by AndreaLK3 - 3
Result of wt103_base
#99 opened by stylelohan - 0
qkv computation
#105 opened by donglixp - 0
what if mems is None?
#104 opened by LindgeW - 2
Clarify why evaluation will be much faster?
#101 opened by cloudygoose - 0
How mem_len affects 1-billion lm experiment result
#102 opened by cmathx - 3
Possible bug in a call?
#96 opened by tridemax - 0
Extending the model for sentiment analysis
#100 opened by CapGOGO - 0
Computing just logits/log prob without getting loss from Adaptive Softmax
#98 opened by tonyhqanguyen - 1
How to speed up the inference
#97 opened by Peony-Wang