issues
search
venkatasg
/
Lil-Bevo
UT Austin's submission to BabyLM Challenge
https://huggingface.co/collections/venkatasg/babylm-653591cdb66f4bf68922873a
MIT License
2
stars
2
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Update README.md
#36
venkatasg
closed
10 months ago
0
final updates
#35
venkatasg
closed
1 year ago
0
updated results
#34
venkatasg
closed
1 year ago
0
Update train_clm.py
#33
juand-r
closed
1 year ago
0
updated results
#32
venkatasg
closed
1 year ago
0
Results updated with short model and ablations
#31
venkatasg
closed
1 year ago
0
final results
#30
venkatasg
closed
1 year ago
0
updated README and spreadsheet. Deleted unnecessary files
#29
venkatasg
closed
1 year ago
0
Update train_clm.py
#28
juand-r
closed
1 year ago
0
Larger tokenizer for 100M dataset
#27
venkatasg
closed
1 year ago
0
Fast tokenizer with mask, cls, padding
#26
venkatasg
closed
1 year ago
0
new Tokenizer, updated results
#25
venkatasg
closed
1 year ago
0
changed name of iter variable to match with huggingface
#24
venkatasg
closed
1 year ago
0
updated README to be clearer
#23
venkatasg
closed
1 year ago
0
music results
#22
venkatasg
closed
1 year ago
0
Update README.md
#21
juand-r
closed
1 year ago
0
fixed bug with abnormally low loss when running training_bevo.py
#20
juand-r
closed
1 year ago
0
updated charts and training script hyperparams
#19
venkatasg
closed
1 year ago
0
MAESTRO results
#18
venkatasg
closed
1 year ago
0
Music pretraining scripts and results
#17
venkatasg
closed
1 year ago
0
Added script for encoder training
#16
venkatasg
closed
1 year ago
0
add tokenizer training wrapper script
#15
alephic
closed
1 year ago
0
Train on synthetically generated data that has some linguistic biases.
#14
venkatasg
closed
1 year ago
1
Train on audio clips of natural sounds
#13
venkatasg
closed
1 year ago
1
Bechmarked decoder models
#12
venkatasg
closed
1 year ago
0
Interpretability of our final model
#11
venkatasg
closed
1 year ago
1
Train on music
#10
venkatasg
closed
1 year ago
3
Contrastive loss
#9
venkatasg
closed
1 year ago
2
Custom models and existing model and training scripts that work towards evaluation
#8
venkatasg
closed
1 year ago
0
script to convert trained model to HF for submission, and data loaded without any stop tokens
#7
venkatasg
closed
1 year ago
0
First working version of LM
#6
venkatasg
closed
1 year ago
0
Get eval pipeline up and running
#5
mahowak
closed
1 year ago
2
Stuff we want to try for loose track
#4
venkatasg
closed
1 year ago
1
Training objectives
#3
venkatasg
closed
1 year ago
0
Training on non-linguistic data
#2
venkatasg
closed
1 year ago
0
Replicate OPT-125M BabyLM benchmark
#1
venkatasg
closed
1 year ago
1