TemporalMesh Transformer: 29.4 PPL at 48% compute — beats Mamba, new open-source architecture
#18 opened about 22 hours ago
by 
vigneshwar234
Request: DOI
#17 opened 5 months ago
by 
cla1r3
Request: DOI
#16 opened 9 months ago
by 
SathiyajithX
Request: DOI
#15 opened over 1 year ago
by 
Atharvab7
Request: DOI
#13 opened over 1 year ago
by 
JK-slone
Attention matrix
1
#12 opened almost 3 years ago
by 
stolosa
Lower precision
#11 opened almost 3 years ago
by 
pipparichter
PEFT LoRA and QLoRA
#10 opened almost 3 years ago
by 
AmelieSchreiber
accessing to embedding layer and generate embeddings step by step
#9 opened about 3 years ago
by 
francescopatane
Understanding vocabulary size
#8 opened about 3 years ago
by 
dannyLCG
how visualize attention matrix
2
#7 opened about 3 years ago
by francescopatane
TorchScript export failed. Maybe related to sequence length cache.
#5 opened over 3 years ago
by 
chenchaozhao
inferring device map for model
#4 opened over 3 years ago
by 
mahdi-b
passing parameters to the underlying model's forward
4
#3 opened over 3 years ago
by mahdi-b