issues
search
helpmefindaname
/
transformer-smaller-training-vocab
Temporary remove unused tokens during training to save ram and speed.
https://helpmefindaname.github.io/transformer-smaller-training-vocab/
MIT License
20
stars
2
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
fix handling of special tokens for tokenizer that have strange buildups
#16
helpmefindaname
opened
1 week ago
0
KeyError: '<unk>'
#15
david-waterworth
opened
2 weeks ago
5
New Tokenizer for mdeberta-v3-base
#14
zynos
opened
1 month ago
1
Increase compability for python 3.12
#13
helpmefindaname
closed
3 months ago
0
Python 3.12 Support
#12
b3n4kh
closed
3 months ago
1
set the vocab size correctly when recreating the full embedding
#11
helpmefindaname
closed
7 months ago
0
Fix version constrain for torch
#10
marcelotrevisani
closed
9 months ago
1
bump python version & poetry version and make datasets an optional dependency
#9
helpmefindaname
closed
10 months ago
0
Sphinx multiversioning
#8
helpmefindaname
closed
1 year ago
0
Add doc page
#7
helpmefindaname
closed
1 year ago
0
fix handling of added special tokens in tokenizers
#6
helpmefindaname
closed
1 year ago
0
add support for torch 2.0.0
#5
helpmefindaname
closed
1 year ago
0
Fix saving of reduced models
#4
helpmefindaname
closed
1 year ago
0
Update Optimizer Parameters
#3
helpmefindaname
closed
1 year ago
0
make dependencies more loose but still working.
#2
helpmefindaname
closed
1 year ago
0
Init
#1
helpmefindaname
closed
1 year ago
0