Adding torch examples

[ncguilbeault]

This PR introduces new examples for the Bonsai.ML.Torch package in bonsai-rx/machinelearning#48. A list of examples and brief summaries are provided below.

Examples

Linear Regression

In this example, the Bonsai.ML.Torch package is used to perform simple linear regression using the ordinary least squares method on simulated data.

Neural Nets - Pretrained Model

This example shows how to run inference using a pretrained convolutional neural network to solve a typical image classification problem. The network is trained to categorize clothing items in the fashion-mnist dataset and perform online inference of new images that were not included in the training dataset. In this particular example, the pretrained weights of the model are loaded from disk.

Neural Nets - Torch Hub

This example shows how a PyTorch model can be downloaded from the PyTorch hub and used to run online inference in Bonsai. For this particular example, the MiDaS model is used to infer the relative inverse depth of images streamed from a camera.

Neural Nets - Training Online

This example shows how a convolutional neural network can be trained online to solve an image classification problem. It demonstrates a typical training procedure that involves performing SGD on the weights of the network as new images and labels are passed to the network. After a certain amount of training, the model can be switch to online inference only mode.

Running On Mono

This example demonstrates how to run the package on Mono. This currently requires a workaround to the DLL import which explicitly throws an error on non-windows OS's running .NET Framework (which is the case for Mono). See this issue: dotnet/TorchSharp#689.

[glopesdev]

Everything looks good, just a few suggestions to smooth out the process. Created a branch dev/torch-examples-torchhub with a suggestion to improve python bootstrapping using uv.

[ncguilbeault]

Thanks for the feedback. I think I've addressed all of the comments from the previous review. Let me know what you think!