issues
search
Guangxuan-Xiao
/
torch-int
This repository contains integer operators on GPUs for PyTorch.
MIT License
181
stars
50
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
fatal error: crypt.h: No such file or directory
#28
rhmaaa
opened
7 months ago
0
is it possible to install torch-int on CUDA version 12.3
#27
ghost
opened
7 months ago
0
what is ellm.tools.quantize_int?
#26
DongHun1995
opened
8 months ago
0
Could not find compiler set in environment variable CUDACXX
#25
Arthur-Ling
opened
11 months ago
2
ZIP does not support timestamps before 1980
#24
moonlightian
closed
11 months ago
1
Undefined symbol when running test_linear_modules.py
#23
jhss
opened
11 months ago
5
Cannot read from remote repo when cloning './torch-int/submodules/cutlass'
#22
FurryMushroom
opened
1 year ago
1
from .._CUDA import (linear_a8_w8_b32_o32, ImportError: attempted relative import with no known parent package
#21
ICE2LAN
opened
1 year ago
5
error: "auto" is not allowed here
#20
oldsweet
opened
1 year ago
1
3 | #include <cutlass/core_io.h> | ^~~~~~~~~~~~~~~~~~~ compilation terminated
#19
baiSongL
opened
1 year ago
2
‼️A100/A800 not working
#18
zhiwei-dong
opened
1 year ago
4
Reduce alpha,beta unnecessary d2h
#17
sleepcoo
opened
1 year ago
0
change all static link library to dynamic so when use nvcr.io/nvidia/pytorch:22.10-py3 container
#16
jackzhou121
closed
1 year ago
0
your environment can never be created
#15
mit10000
opened
1 year ago
2
can not install by setup.py because /ld: cannot find -lcublas_static
#14
meicale
opened
1 year ago
5
Add extra include dir, needed with cuda11.8
#13
CL-Shang
opened
1 year ago
0
Tests `test_linear_shape.py` fails
#12
FabianSchuetze
opened
1 year ago
0
编译报错,是什么原因啊
#11
o-github-o
opened
1 year ago
0
ModuleNotFoundError: No module named 'torch_int._CUDA'
#10
aakejiang
closed
1 year ago
8
Update .gitmodules
#9
ZhengHSI
closed
1 year ago
0
error: "auto" is not allowed here
#8
ChanChiChoi
closed
1 year ago
0
A version conflict of cutlass & torch-int
#7
Ther-nullptr
opened
1 year ago
1
Make the dequant type generic to support other input types eg bfloat16
#6
erichan1
opened
1 year ago
0
file benchmark/bench_linear.sh is broken
#5
pommedeterresautee
opened
1 year ago
1
Add cmake dependency in Readme
#4
mickaelseznec
closed
1 year ago
0
CUTLASS submodule URL might not work
#3
mickaelseznec
opened
1 year ago
1
Support cuda architecture sm70 and sm75
#2
merrymercy
closed
1 year ago
1
adding support for GPT-J
#1
ImanHosseini
opened
1 year ago
12