# mdgrad
A small autograd engine that implements backpropagation (reverse-mode autodiff). Heavily inspired by karpathy's [micrograd](https://github.com/karpathy/micrograd/tree/master), and extended to support operations on tensors instead of scalars. Includes a small neural network api for building and training neural networks. Has a PyTorch-like API.
Hopefully useful as an educational resource.
## Installation
``` bash
pip install mdgrad
```
## Example Usage
A dumb example showing supported operations
```python
import mdgrad
import mdgrad.nn as nn
a = 3 * mdgrad.randn(3, 2)
b = mdgrad.ones(shape=(2, 2))
c = a @ b
d = c * 3 / 2
e = d ** 2
f = e.sum()
print(f.data)
f.backward()
print(a.grad)
```
An example showing how to define and run a neural network. See the files in `examples/` for more details on building and training models.
```python
import mdgrad
import mdgrad.nn as nn
# Define the model and loss function
model = nn.Sequential(
nn.Linear(2, 20),
nn.ReLU(),
nn.Linear(20, 50),
nn.ReLU(),
nn.Linear(50, 15),
nn.ReLU(),
nn.Linear(15, 1),
nn.Sigmoid()
)
loss_fn = nn.MSELoss()
# Create dummy data
X = mdgrad.randn(100, 2)
target = mdgrad.randn(100, 1)
# Compute output and loss
out = model(X)
loss = loss_fn(out, target)
# Compute gradients of parameters
loss.backward()
```
Raw data
{
"_id": null,
"home_page": null,
"name": "mdgrad",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.8",
"maintainer_email": null,
"keywords": "python, tensors, neural networks, automatic differentiation",
"author": "Jordan Madden",
"author_email": "<jordanmadden285@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/53/ec/42b7dfba183465ea519c7f7407b39a53cb7b9a4215cfacca62da07cddc63/mdgrad-0.3.tar.gz",
"platform": null,
"description": "# mdgrad\n\nA small autograd engine that implements backpropagation (reverse-mode autodiff). Heavily inspired by karpathy's [micrograd](https://github.com/karpathy/micrograd/tree/master), and extended to support operations on tensors instead of scalars. Includes a small neural network api for building and training neural networks. Has a PyTorch-like API.\n\nHopefully useful as an educational resource.\n\n## Installation\n\n``` bash\npip install mdgrad\n```\n\n## Example Usage\n\nA dumb example showing supported operations\n\n```python\n\nimport mdgrad\nimport mdgrad.nn as nn\n\na = 3 * mdgrad.randn(3, 2)\nb = mdgrad.ones(shape=(2, 2))\nc = a @ b\nd = c * 3 / 2\ne = d ** 2\nf = e.sum()\nprint(f.data) \nf.backward()\nprint(a.grad) \n```\n\nAn example showing how to define and run a neural network. See the files in `examples/` for more details on building and training models.\n\n```python\n\nimport mdgrad\nimport mdgrad.nn as nn\n\n# Define the model and loss function\nmodel = nn.Sequential(\n nn.Linear(2, 20),\n nn.ReLU(),\n nn.Linear(20, 50), \n nn.ReLU(),\n nn.Linear(50, 15),\n nn.ReLU(),\n nn.Linear(15, 1),\n nn.Sigmoid()\n)\nloss_fn = nn.MSELoss()\n\n# Create dummy data\nX = mdgrad.randn(100, 2)\ntarget = mdgrad.randn(100, 1)\n\n# Compute output and loss\nout = model(X)\nloss = loss_fn(out, target)\n\n# Compute gradients of parameters\nloss.backward()\n```\n",
"bugtrack_url": null,
"license": null,
"summary": "Tensor-based autdiff engine and neural network API",
"version": "0.3",
"project_urls": null,
"split_keywords": [
"python",
" tensors",
" neural networks",
" automatic differentiation"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "72d5c51cc8724e7355b7720efbdca99470073609b40e64013a6fbda48c1d48d8",
"md5": "9aa352c2b2517be67d4e9e0a79583027",
"sha256": "fad050d5ef563d6b5348d276420c394aacac7abe2a7e14a1cd1ebfefc01f0d52"
},
"downloads": -1,
"filename": "mdgrad-0.3-py3-none-any.whl",
"has_sig": false,
"md5_digest": "9aa352c2b2517be67d4e9e0a79583027",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.8",
"size": 12018,
"upload_time": "2024-07-13T17:55:19",
"upload_time_iso_8601": "2024-07-13T17:55:19.964692Z",
"url": "https://files.pythonhosted.org/packages/72/d5/c51cc8724e7355b7720efbdca99470073609b40e64013a6fbda48c1d48d8/mdgrad-0.3-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "53ec42b7dfba183465ea519c7f7407b39a53cb7b9a4215cfacca62da07cddc63",
"md5": "5cc438e9c496ad59e0b6b5a4f0854742",
"sha256": "f4457a9d014a65e7f5972d147db4d0edc248727d316c3cba7f5575d7e4a548fc"
},
"downloads": -1,
"filename": "mdgrad-0.3.tar.gz",
"has_sig": false,
"md5_digest": "5cc438e9c496ad59e0b6b5a4f0854742",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.8",
"size": 11792,
"upload_time": "2024-07-13T17:55:21",
"upload_time_iso_8601": "2024-07-13T17:55:21.571212Z",
"url": "https://files.pythonhosted.org/packages/53/ec/42b7dfba183465ea519c7f7407b39a53cb7b9a4215cfacca62da07cddc63/mdgrad-0.3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-07-13 17:55:21",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "mdgrad"
}
JSON