-
Does anyone have any idea about loading a custom dataset for tracking scenarios?
I managed to load a dataset for the detection scenario by modifying "python src/generate_coco_from_mot.py," but I a…
-
Hi. It's a very nice website. But I have some questions about the data.
On the website,
"read 1000,000 bytes sequentially from memory is 3us", so the memory bandwidth should be 1MB/3us = 333GB/s.
…
-
# What
The example IDs are globally registered, thus creating conflicts with newly added example IDs in `data/examples.txt`. This conflict makes the build fail and abort the process. Thus it's not po…
-
will be good to have option to delay writing of all of metadata (Directory and FAT entries).
So it will be possible to sequential write data of many files without interrupting to write its metadata.…
-
Truly a nice work!
Video data often necessitates incorporating the temporal dimension into the network architecture to account for its sequential frames. However, DINO v2 currently lacks provisions f…
-
Currently, we have three data partitioning levels: shards, segments, and docs. By default, we use segments. We don't expose these options to users because the overhead of doc partitioning can be expen…
-
### 🐛 Describe the bug
```python
import torch.export
m = torch.nn.Sequential(
torch.nn.ReLU(),
)
x = torch.ones(3)
r1 = m(x)
sym = torch.export.unflatten(torch.export.export(m, (x,)))
r2 …
-
Very often in ML people that use sample want to do the following: shuffle the entire data and then start reading them sequentially. Once all data is read, shuffle again and read sequentially and so on…
-
### Deep Learning Simplified Repository (Proposing new issue)
:red_circle: **Project Title** : Facial Skin Diseases Classification using DL
:red_circle: **Aim** : The aim of this project is to class…
-
**Is your feature request related to a problem? Please describe.**
The PyT and TF Dataloader support padding list (sparse) features to the right, which means that shorter list sequences will be compl…