issues
search
lucidrains
/
flash-attention-jax
Implementation of Flash Attention in Jax
MIT License
187
stars
23
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
fixed bug that occurred if query/key and value dimensions were different
#15
ihh
opened
6 months ago
0
Reshape error in causal_flash_attention when sequence length is not a multiple of 1024
#14
enolan
opened
8 months ago
0
Return same dtype as is passed in in qkv
#13
enolan
opened
1 year ago
0
Slower than non-flash attention
#12
aniquetahir
closed
1 year ago
1
Multi-head causal flash attention support?
#11
sh0416
closed
1 year ago
8
Question about calculation of Q and transpose(K).
#10
jaes77
opened
1 year ago
0
Performance benchmarks?
#9
imoneoi
opened
1 year ago
21
refactor custom_vjp use for type agreement
#8
mattjj
closed
1 year ago
1
fix compatibility with jax transformations
#7
GallagherCommaJack
opened
1 year ago
28
support for per-head scales for cosine sim attention
#6
GallagherCommaJack
opened
1 year ago
6
more general mask support
#5
GallagherCommaJack
opened
1 year ago
1
batch & multihead support?
#4
GallagherCommaJack
closed
1 year ago
3
can I work on making a flax attention function out of this repository?
#3
MiladInk
opened
1 year ago
1
Online Softmax from FlashAttention
#2
jenkspt
closed
1 year ago
2
Add JIT to flash attention
#1
VHellendoorn
closed
2 years ago
4