issues
search
tairov
/
llama2.mojo
Inference Llama 2 in one file of pure 🔥
https://www.modular.com/blog/community-spotlight-how-i-built-llama2-by-aydyn-tairov
MIT License
2.09k
stars
139
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
#+-13-20% speed up making steps as an alias, and using it in a range
#40
rd4com
closed
11 months ago
2
Optimize transformer loops
#39
mikowals
closed
11 months ago
7
replace Matrix with Tensor
#38
mikowals
closed
11 months ago
6
Getting very strange response when trying the second example in README.md
#37
wongni
closed
11 months ago
2
Killed
#36
Tsunami014
closed
11 months ago
2
Remove Falcon style ROPE
#35
magician-blue
closed
11 months ago
10
Remove falcon style rope
#34
magician-blue
closed
11 months ago
2
it is taking python 3.6 insted of 3.9.18
#33
y3-rawat
closed
11 months ago
3
llama2.c Tinyllama1.1B supported.
#32
magician-blue
closed
11 months ago
8
In some case, we haven't make full use of threads
#31
magician-blue
closed
10 months ago
8
Better tokenizer and tinyllama-1.1B supported
#30
magician-blue
closed
11 months ago
4
Add boundary checks to FileBuf
#29
mikowals
closed
11 months ago
2
Question about models
#28
Ezyweb-uk
closed
11 months ago
2
Make the tokenizer better
#27
magician-blue
closed
11 months ago
33
binary search to get prompt tokens
#26
mikowals
closed
11 months ago
3
remove overwrite of vocab_scores
#25
mikowals
closed
11 months ago
4
Update llama2.mojo from MHA to GQA
#24
magician-blue
closed
11 months ago
23
Is llama2 a group query attention or multi head attention?
#23
magician-blue
closed
11 months ago
3
How it relates to LLAMA?
#22
alexcardo
closed
11 months ago
1
Stop generating when EOS BOS appear
#21
tic-top
closed
11 months ago
4
[wip] Tensor refactoring attempt
#20
tairov
closed
11 months ago
5
llama2.mojo
#19
hemabatchu
closed
11 months ago
0
Mojo: error: execution exited with a non-zere result:1
#18
Cyria7
closed
11 months ago
1
Can a RestAPI be built
#17
StudyingLover
closed
11 months ago
2
[wip] Apply more parallelize
#16
tairov
closed
11 months ago
0
vectorize llama methods
#15
tairov
closed
11 months ago
1
Add extra params to Gradio UI
#14
radames
closed
11 months ago
0
Segmentation falut error when it is built as binary
#13
3DAlgoLab
closed
11 months ago
8
mojo llama2.mojo error
#12
mathpopo
closed
11 months ago
1
Prompt input?
#11
radames
closed
11 months ago
6
Simple Gradio UI + Dockerfile
#10
radames
closed
11 months ago
4
Woraround parallelize() issues and default to parallel matmul
#9
itramble
closed
11 months ago
4
question:Are these speed comparisons all in CPU mode? Can we add a comparison with GPU?
#8
xiezhipeng-git
closed
11 months ago
1
How to create my own tokenizer.bin?
#7
hungtooc
closed
11 months ago
3
How to use mojo playground to run this model?
#6
yhyu13
closed
11 months ago
1
Increase number of SIMD registers used to speed-up the execution
#5
VMois
closed
11 months ago
9
Where is 'tokenizer.bin'?
#4
shi3z
closed
11 months ago
1
add tokenizer from llama2.py
#3
g1y5x3
closed
11 months ago
1
Error executing mojo llama2.mojo
#2
mpage38
closed
11 months ago
11
Use simd value directly in matmul
#1
davors72
closed
11 months ago
1
Previous