issues
search
learning-at-home
/
lean_transformer
Memory-efficient transformer. Work in progress.
MIT License
19
stars
3
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
attention cache
#13
xtinkt
opened
2 years ago
0
add attn cache
#12
xtinkt
closed
2 years ago
1
Research diary: reversibles
#11
justheuristic
opened
2 years ago
1
multihead attention version with loop by batch dimension to reduce memory usage
#10
krunt
opened
2 years ago
2
First and second predictions yield slightly different results when both jit and dropout are enabled
#9
justheuristic
opened
2 years ago
0
Check out FMHA
#8
justheuristic
opened
2 years ago
1
Memory-efficient attention
#7
justheuristic
opened
2 years ago
6
Roadmap
#6
justheuristic
opened
2 years ago
0
Inference / LeanGPT.generate
#5
justheuristic
opened
2 years ago
1
Potential optimizations
#4
justheuristic
opened
2 years ago
0
CI suite
#3
justheuristic
opened
2 years ago
1
Vision transformers
#2
justheuristic
opened
2 years ago
0
Additional tests
#1
justheuristic
opened
2 years ago
0