Computer Vision models and training
Project description
Quickvision
- Faster Computer Vision.
Install Quickvision
- Install directly from GitHub. Very soon it will be available over PyPi.
pip install -q git+https://github.com/Quick-AI/quickvision.git
What is Quickvision?
- Quickvision makes Computer Vision tasks much faster and easier with PyTorch.
It provides: -
- Easy to use torch native API, for
fit()
,train_step()
,val_step()
of models. - Easily customizable and configurable models with various backbones.
- A complete torch native interface. All models are
nn.Module
all the training APIs are optional and not binded to models. - A lightning API which helps to accelerate training over multiple GPUs, TPUs.
- A datasets API to common data format very easily and quickly to torch formats.
- A minimal package, with very low dependencies.
- Train your models faster. Quickvision has already implmented the long learning in torch.
Quickvision is just Torch!!
- Quickvision does not make you learn a new library. If you know PyTorch you are good to go!!!
- Quickvision does not abstract any code from torch, nor implements any custom classes over it.
- It keeps the data format in
Tensor
only. You don't need to convert it.
Do you want just a model with some backbone configuration?
- Use model made by us. It's just a
nn.Module
which has Tensors only Input and Output format. - Quickvision provides reference scripts too for training it!
Do you want to train your model but not write lengthy loops?
- Just use our training methods such as
fit()
,train_step()
,val_step()
.
Do you want multi GPU training but worried about model configuration?
- Just subclass the PyTorch Lightning model!
- Implement the
train_step
,val_step
.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
quickvision-0.1.0.tar.gz
(34.4 kB
view hashes)
Built Distribution
Close
Hashes for quickvision-0.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | bcf503ef6d78e6f80f02ce9e724351e9cb3c3edf31d74ef30e6fc10520ff41c2 |
|
MD5 | 32da9ef3f3bf7eebfa2dd9b24235935a |
|
BLAKE2b-256 | 966ccef155c3ecb775d2a27aed5fca199baa61f4995e747bb72c576ab78b879e |