A Rust crate for building and training neural networks. tinygrad provides a simple interface for defining tensors, performing forward and backward passes, and implementing basic operations such as dot products and summation.
Get started with the tinygrad library by following these simple steps:
- Install the
tinygradcrate by adding the following line to yourCargo.tomlfile:
[dependencies]
tinygrad = "0.1.0"- Use the
TensorandForwardBackwardtraits to create and work with tensors:
use ndarray::{array, Array1};
use tinygrad::{Tensor, Context, TensorTrait};
// Create a tensor
let value = array![1.0, 2.0, 3.0];
let tensor = Tensor::new(value);
// Perform forward and backward passes
let mut ctx = Context::new();
let result = tensor.forward(&mut ctx, vec![tensor.get_value()]);
tensor.backward(&mut ctx, array![1.0, 1.0, 1.0].view());- Implement custom operations by defining structs that implement the
ForwardBackwardtrait:
use ndarray::ArrayView1;
use tinygrad::{ForwardBackward, Context, TensorTrait};
// Example operation: Dot product
struct Dot;
impl ForwardBackward for Dot {
fn forward(&self, _ctx: &mut Context, inputs: Vec<ArrayView1<f64>>) -> f64 {
let input = &inputs[0];
let weight = &inputs[1];
input.dot(weight)
}
fn backward(&self, ctx: &mut Context, grad_output: ArrayView1<f64>) {
// Implement backward pass
// ...
}
}use ndarray::{array, Array1};
use tinygrad::{Tensor, Context, TensorTrait};
fn main() {
let input = array![1.0, 2.0, 3.0];
let weight = array![4.0, 5.0, 6.0];
let input_tensor = Box::new(Tensor::new(input));
let weight_tensor = Box::new(Tensor::new(weight));
let dot_fn = Dot;
let mut ctx = Context::new();
let inputs = vec![
input_tensor.get_value(),
weight_tensor.get_value(),
];
let output = dot_fn.forward(&mut ctx, inputs);
println!("Dot product: {:?}", output);
let grad_output = array![1.0, 1.0, 1.0];
dot_fn.backward(&mut ctx, grad_output.view());
let grad_input = &input_tensor.grad.clone();
let grad_weight = &weight_tensor.grad.clone();
println!("Gradient for input: {:?}", grad_input);
println!("Gradient for weight: {:?}", grad_weight);
}Run tests for the tinygrad crate using:
cargo testYou can access the source code for the tinygrad crate on GitHub.
Contributions and feedback are welcome! If you'd like to contribute, report an issue, or suggest an enhancement, please engage with the project on GitHub. Your contributions help improve this crate for the community.
Full documentation for tinygrad is available on docs.rs.
This project is licensed under the MIT License.