-
I want each batch to always consist of images from two different datasets.
I tried to define two prefetchers or define two dataloader, but either way I stopped there when executing the following stat…
-
If bg_num < batch_size, there is a error like title.
And it will always one image, when bg_num = 1, batch_size = 1.
-
### 🐛 Describe the bug
When running prefetch on multiple branches of a forked datapipe, it is possible to trigger a race condition.
```python
import time
import torchdata.datapipes as dp
from…
-
Hi, I saw this fork from the original stabilizer repo and is very interested in it. I wanted to know how much a difference does this make for modern hardware with much larger cache size, more associat…
-
I tried to reproduce the results in this article.
I know that uarch-bench can use perf, but since I always get jevents errors, I made a simple implementation of this part of uarch-bench, and the co…
-
Just found an issue on K230 when doing some auto-vectorization tests on https://github.com/UoB-HPC/TSVC_2.
The vectorized `s1115` is like:
```asm
.LBB9_7: # %vector.p…
-
When running long sequences with the APS2 if the waveform data cross a cache line the prefetcher can't keep up. The experiment proceeds, but the data are random.
Verified this for example with the …
-
I have PyTorch v1.2 and Apex installed. I was able to get opt_level 00 and 03 working. But, opt_level 01 option isn't working. From the debugging I've done, looks like it's crashing at ` prefetcher…
-
When I use 'GPUS_PER_NODE=1 ./tools/run_dist_launch.sh 1 ./configs/r50_deformable_detr.sh' to try to train on the coco dataset, The following error occurred:
Start training
Traceback (most recent …
-
## Decode 128
We are aiming to hit 20 t/s/u end to end for Llama3 decode on t3k. These are the issues left.
![image](https://github.com/user-attachments/assets/ce4a7190-f49d-431b-8f18-79ebb130e36c…