wkcn / Mobulaop
Licence: mit
A Simple & Flexible Cross Framework Operators Toolkit
Stars: ✭ 161
Programming Languages
Projects that are alternatives of or similar to Mobulaop
Tensorly
TensorLy: Tensor Learning in Python.
Stars: ✭ 977 (+506.83%)
Mutual labels: mxnet, cupy, numpy
NPY-for-Fortran
A FORTRAN module to write Numpy's *.npy and *.npz files
Stars: ✭ 30 (-81.37%)
Mutual labels: numpy, scientific-computing
Xshinnosuke
Deep learning framework realized by Numpy purely, supports for both Dynamic Graph and Static Graph with GPU acceleration
Stars: ✭ 291 (+80.75%)
Mutual labels: cupy, numpy
SciCompforChemists
Scientific Computing for Chemists text for teaching basic computing skills to chemistry students using Python, Jupyter notebooks, and the SciPy stack. This text makes use of a variety of packages including NumPy, SciPy, matplotlib, pandas, seaborn, NMRglue, SymPy, scikit-image, and scikit-learn.
Stars: ✭ 65 (-59.63%)
Mutual labels: numpy, scientific-computing
Datacube Core
Open Data Cube analyses continental scale Earth Observation data through time
Stars: ✭ 285 (+77.02%)
Mutual labels: numpy, scientific-computing
Einops
Deep learning operations reinvented (for pytorch, tensorflow, jax and others)
Stars: ✭ 4,022 (+2398.14%)
Mutual labels: cupy, numpy
Chainer
A flexible framework of neural networks for deep learning
Stars: ✭ 5,656 (+3413.04%)
Mutual labels: cupy, numpy
Eliot
Eliot: the logging system that tells you *why* it happened
Stars: ✭ 874 (+442.86%)
Mutual labels: numpy, scientific-computing
PyRates
Open-source, graph-based Python code generator and analysis toolbox for dynamical systems (pre-implemented and custom models). Most pre-implemented models belong to the family of neural population models.
Stars: ✭ 33 (-79.5%)
Mutual labels: numpy, scientific-computing
Ivy
The templated deep learning framework, enabling framework-agnostic functions, layers and libraries.
Stars: ✭ 118 (-26.71%)
Mutual labels: mxnet, numpy
robot
Functions and classes for gradient-based robot motion planning, written in Ivy.
Stars: ✭ 29 (-81.99%)
Mutual labels: mxnet, numpy
Python-Matematica
Explorando aspectos fundamentais da matemática com Python e Jupyter
Stars: ✭ 41 (-74.53%)
Mutual labels: numpy, scientific-computing
Pynvvl
A Python wrapper of NVIDIA Video Loader (NVVL) with CuPy for fast video loading with Python
Stars: ✭ 95 (-40.99%)
Mutual labels: cupy, numpy
Numcpp
C++ implementation of the Python Numpy library
Stars: ✭ 2,031 (+1161.49%)
Mutual labels: numpy, scientific-computing
Pyeco
python implementation of efficient convolution operators for tracking
Stars: ✭ 150 (-6.83%)
Mutual labels: mxnet
MobulaOP
Linux | Windows | Coverage | Badge |
---|---|---|---|
What is it?
MobulaOP is a simple and flexible cross framework operators toolkit.
You can write custom operators by Python/C++/C/CUDA/HIP/TVM without rebuilding deep learning framework from source.
How to use it?
[中文教程]
[Tutorial]
- Add an addition operator [Code]
import mobula
@mobula.op.register
class MyFirstOP:
def forward(self, x, y):
return x + y
def backward(self, dy):
return [dy, dy]
def infer_shape(self, in_shape):
assert in_shape[0] == in_shape[1]
return in_shape, [in_shape[0]]
# MXNet
import mxnet as mx
a = mx.nd.array([1, 2, 3])
b = mx.nd.array([4, 5, 6])
c = MyFirstOP(a, b)
print (c) # [5, 7, 9]
# PyTorch
import torch
a = torch.tensor([1, 2, 3])
b = torch.tensor([4, 5, 6])
c = MyFirstOP(a, b)
print (c) # [5, 7, 9]
# NumPy
import numpy as np
a = np.array([1, 2, 3])
b = np.array([4, 5, 6])
op = MyFirstOP[np.ndarray]()
c = op(a, b)
print (c) # [5, 7, 9]
# CuPy
import cupy as cp
a = cp.array([1, 2, 3])
b = cp.array([4, 5, 6])
op = MyFirstOP[cp.ndarray]()
c = op(a, b)
print(c) # [5, 7, 9]
- Use custom operators without rebuilding the source of deep learning framework [Code]
# Use ROIAlign operator
import mxnet as mx
import numpy as np
import mobula
# Load ROIAlign Module
mobula.op.load('ROIAlign')
ctx = mx.cpu(0)
dtype = np.float32
N, C, H, W = 2, 3, 4, 4
data = mx.nd.array(np.arange(N*C*H*W).astype(dtype).reshape((N,C,H,W)))
rois = mx.nd.array(np.array([[0, 1, 1, 3, 3]], dtype = dtype))
data.attach_grad()
with mx.autograd.record():
# mx.nd.NDArray and mx.sym.Symbol are both available as the inputs.
output = mobula.op.ROIAlign(data = data, rois = rois, pooled_size = (2,2), spatial_scale = 1.0, sampling_ratio = 1)
print (output.asnumpy(), data.grad.asnumpy())
- Import Custom C++ Operator Dynamically [Code]
import mobula
# Import Custom Operator Dynamically
mobula.op.load('./AdditionOP')
import mxnet as mx
a = mx.nd.array([1,2,3])
b = mx.nd.array([4,5,6])
c = mobula.op.AdditionOP(a, b)
print ('a + b = c \n {} + {} = {}'.format(a.asnumpy(), b.asnumpy(), c.asnumpy()))
How to get it?
# Clone the project
git clone https://github.com/wkcn/MobulaOP
# Enter the directory
cd MobulaOP
# Install MobulaOP
pip install -v -e .
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].