DALI
SlowFast
Our great sponsors
| DALI | SlowFast | |
|---|---|---|
| 5 | 7 | |
| 4,914 | 6,273 | |
| 2.1% | 2.0% | |
| 9.6 | 5.1 | |
| 2 days ago | 4 months ago | |
| C++ | Python | |
| Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
DALI
-
[D] Will data augmentations work faster on TPUs?
Another option is DALI https://github.com/NVIDIA/DALI For my project while training EfficientNet2, it was a game changer. But it a way harder to implement in code than TorchVision or Kornia.
-
DirectStorage - Loading data to GPU *directly* from the SSD drive, almost without using CPU
Check out https://github.com/nvidia/DALI
-
mmap_ninja: Speedup your training dramatically by using memory-mapped files for your dataset
Small question if you are using GPU: How to this compare to GPUDirect Storage from Nvidia? can you have even more speedup by using both? I never toy with it, but the DALI project from Nvidia seem to tackle the same data loading problem.
- [D] Efficiently loading videos in PyTorch without extracting frames
SlowFast
-
Counting number of times a behavior is repeated with Computer Vision: How hard is this?
For X3D check : Meta's Repo
-
Questions about some video-related concepts such as sampling rate and frame rate
However, when I am reading the slowfast code （https://github.com/facebookresearch/SlowFast/blob/main/slowfast/config/defaults.py）
-
[D] Efficiently loading videos in PyTorch without extracting frames
Try using ffmpeg. They use av Python bindings in this repository: https://github.com/facebookresearch/SlowFast/tree/main/slowfast/datasets
-
Facebook AI Introduces Multiscale Vision Transformers (MViT), A Transformer Architecture For Representation Learning From Visual Data
Code for https://arxiv.org/abs/2104.11227 found: https://github.com/facebookresearch/SlowFast
Github: https://github.com/facebookresearch/SlowFast
-
[R] Facebook AI Conducts Large-Scale Study on Unsupervised Spatiotemporal Representation Learning
Super interesting! Has anyone managed to find the code for this - the official repo doesn't seem to have the models yet https://github.com/facebookresearch/SlowFast
-
When TimeSFormer repo? [action recognition]
facebook’s TimeSFormer / “is space-time attention all you need for video understanding?” Came out in feb. There is no official repo (yet). Maybe they will add it to https://github.com/facebookresearch/SlowFast
What are some alternatives?
Pytorch - Tensors and Dynamic neural networks in Python with strong GPU acceleration
vision - Datasets, Transforms and Models specific to Computer Vision
Blurry - Blurry is an easy blur library for Android
imutils - A series of convenience functions to make basic image processing operations such as translation, rotation, resizing, skeletonization, and displaying Matplotlib images easier with OpenCV and Python.
executorch - On-device AI across mobile, embedded and edge for PyTorch
DREAMPlace - Deep learning toolkit-enabled VLSI placement
MegEngine - MegEngine 是一个快速、可拓展、易于使用且支持自动求导的深度学习框架
ocaml-torch - OCaml bindings for PyTorch
Image Steganography - ✔️ Hide a secret message in an image
soft_nms - PyTorch implementation of soft-nms
Louvre - A small customizable library useful to handle an gallery image pick action built-in your app. :sunrise_over_mountains::stars: