ryujaehun pytorch-gpu-benchmark issues

ryujaehun / pytorch-gpu-benchmark

Using the famous cnn model in Pytorch, we run benchmarks on various gpu.

MIT License

224 stars 85 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Added 7900XTX results with lamikr's support for AMD cards

#31 eitch opened 2 months ago
0
plot.ipynb does not work for parsing new results with pytorch 2.30

#30 lamikr opened 3 months ago
0
Test.sh needs changes to allow running tests also for AMD gpus

#29 lamikr opened 3 months ago
0
AMD support to test.sh and AMD RX 6800 results

#28 lamikr opened 3 months ago
1
When running benchmark_models.py, show error messages"No module named 'pandas'"

#27 Oracle-Chen opened 5 months ago
0
Added RTX 5000 Ada Laptop performances

#26 dot-Eagle96 opened 5 months ago
0
Add 6800XT results

#25 ToughStyle opened 6 months ago
0
Added 4060 Ti 16GB results

#24 adorobis closed 6 months ago
0
why did this benchmark blow away my linux partition?

#23 achillez closed 1 year ago
1
bugfix , support for torch 2.0

#22 hellojixian closed 1 year ago
1
TypeError: __call__() got an unexpected keyword argument 'pretrained'

#21 joshhu opened 2 years ago
4
adding rtx 3060 results

#20 zachcoleman closed 2 years ago
0
RuntimeError: miopenStatusUnknownError

#19 Bengt opened 2 years ago
0
error in type mobilenet_v2

#18 jjziets closed 3 years ago
5
GTX 2080 Ti Performance [Windows 10]

#17 arstropica closed 3 years ago
1
cannot support ROCM

#16 znsoftm closed 3 years ago
3
titan RTX

#15 ryujaehun closed 3 years ago
0
results for nvidia A100

#14 kirk86 closed 3 years ago
2
poly.ipynb ValueError

#13 sevaroy closed 3 years ago
1
Incorrect help information for arguments: folder

#12 yqtianust closed 4 years ago
3
test rtx2080ti on windows10

#11 olixu closed 4 years ago
0
resnet with batchsize=12?

#10 twmht closed 4 years ago
1
The performance of cuda on windows10 VS Linux

#9 olixu closed 4 years ago
7
If the batch size becomes smaller, it would occurs because the other part is overloaded rather than the operation.

#8 olixu closed 4 years ago
0
have you any interest to try larger batch size, such as 64 or 128?

#7 qixiang109 closed 4 years ago
1
Have you tried batch_size 1 for fp32 and fp16?

#6 Edwardmark closed 4 years ago
1
RuntimeError: CUDA out of memory. Tried to allocate 32.00 MiB (GPU 0; 7.79 GiB total capacity; 6.64 GiB already allocated; 21.00 MiB free; 150.86 MiB cached) .

#5 M-Abdallah closed 4 years ago
4
Fix bar plot layout, add a new figure

#4 elombardi2 closed 5 years ago
0
My 1080 GTX Ti is 50% slower than your benchmark

#3 NikEyX closed 4 years ago
5
Update typo

#2 johmathe closed 5 years ago
0
Summary BenchMark

#1 ryujaehun closed 4 years ago
0