pbloem/former issues and pull requests

#40 - class SelfAttentionRelative(nn.Module):

Issue - State: open - Opened by EltonQm 2 months ago - 1 comment

#39 - def sample_sequence

Issue - State: open - Opened by EltonQm 2 months ago

#38 - compute_compression func returns total bits (not bits per byte)

Issue - State: open - Opened by diwert-ai 2 months ago

#37 - How to tokenize a testing phrase

Issue - State: open - Opened by edgarmg91 about 1 year ago

#36 - Why the use of log_softmax ?

Issue - State: closed - Opened by dsantiago about 1 year ago - 1 comment

#35 - No Field in torchtext.data

Issue - State: open - Opened by dtdo90 over 1 year ago - 1 comment

#34 - Calculation of memory required in "Going big"

Issue - State: closed - Opened by oleksandrasaskia over 1 year ago - 2 comments

#33 - Dev ajay

Pull Request - State: closed - Opened by ajay-usfca about 2 years ago

#32 - Benntuecon/issue2

Pull Request - State: closed - Opened by benntuecon about 2 years ago

#31 - Version1 able to extract

Pull Request - State: closed - Opened by benntuecon about 2 years ago

#30 - Question about k × hk weight matrices

Issue - State: closed - Opened by ramesaliyev about 2 years ago - 1 comment

#29 - improve performance in deeper network using your multi head attention code

Issue - State: closed - Opened by fatemehniknezhad over 2 years ago - 3 comments

#28 - Hi,question about sliced-up version self-attention

Issue - State: open - Opened by Faith-Uchiha over 2 years ago - 1 comment

#27 - conda installation failed

Issue - State: closed - Opened by bsun0802 about 3 years ago - 1 comment

#26 - Blog is down

Issue - State: closed - Opened by JJGO over 3 years ago - 1 comment

#25 - Module import error

Issue - State: open - Opened by Cybernetic1 over 3 years ago - 1 comment

#24 - ImportError: attempted relative import with no known parent package

Issue - State: closed - Opened by RubenRidderstrom over 3 years ago - 8 comments

#23 - Understanding multiheaded attention.

Issue - State: closed - Opened by anorak94 over 3 years ago

#22 - High compute time, what is a reasonable generator model to get somewhat good results to play with?

Issue - State: closed - Opened by jplasser almost 4 years ago - 1 comment

#21 - import SelfAttentionWide (2x) -> import SelfAttentionNarrow, SelfAtte…

Pull Request - State: closed - Opened by bquast almost 4 years ago - 1 comment

#20 - data.Field no longer supported in torchtext

Issue - State: closed - Opened by jvdburgt almost 4 years ago - 1 comment

#19 - slide 50, v and q mixed up

Issue - State: closed - Opened by ulf1 almost 4 years ago - 1 comment

#18 - token_embedding for non-text sequences

Issue - State: closed - Opened by StolkArjen about 4 years ago - 6 comments

#17 - Weight matrix dimensions after transofrmation

Issue - State: closed - Opened by SrinjaySarkar about 4 years ago - 1 comment

#16 - Update init.py

Pull Request - State: closed - Opened by SubhashPadala about 4 years ago

#15 - How can we visualize the self-attention map

Issue - State: closed - Opened by amiltonwong over 4 years ago - 1 comment

#14 - Using the trained model

Issue - State: closed - Opened by ShivanshuPurohit over 4 years ago - 1 comment

#13 - Is narrow attention implemented correctly?

Issue - State: closed - Opened by TheGrayFrost over 4 years ago - 8 comments
Labels: bug

#12 - Accuracies on the examples

Issue - State: open - Opened by sf-wind about 5 years ago - 1 comment

#11 - Question about the tool used to make the figures in the blog post

Issue - State: open - Opened by yassouali about 5 years ago - 2 comments

#10 - Issue with masking

Issue - State: closed - Opened by mc-robinson about 5 years ago - 1 comment

#9 - ModuleNotFoundError: No module named 'past'

Issue - State: closed - Opened by jjong2ya about 5 years ago - 5 comments

#8 - Comparing SelfAttention classes

Issue - State: closed - Opened by sidneyaraujomelo over 5 years ago - 2 comments

#7 - AttributeError: module 'torch' has no attribute 'triu_indices'

Issue - State: closed - Opened by xxxxyan over 5 years ago - 5 comments

#6 - Masking done for the upper or lower triangle?

Issue - State: closed - Opened by esvhd over 5 years ago - 1 comment

#5 - Why is dividing by e**(1/4) for both keys and queries more memory efficient?

Issue - State: open - Opened by esvhd over 5 years ago - 4 comments

#4 - Einsum to avoid transpose and reshape

Issue - State: closed - Opened by azzever over 5 years ago - 3 comments

#3 - add conda env

Pull Request - State: closed - Opened by koen-dejonghe over 5 years ago - 1 comment

#2 - Weight scaling should be for keys not values?

Issue - State: closed - Opened by esvhd over 5 years ago - 2 comments

#1 - ModuleNotFoundError: No module named 'past'

Issue - State: closed - Opened by koen-dejonghe over 5 years ago - 7 comments

GitHub / pbloem/former issues and pull requests