search

TRAIS-Lab / dattri

`dattri` is a PyTorch library for developing, benchmarking, and deploying efficient data attribution algorithms.
https://trais-lab.github.io/dattri/
MIT License
28 stars 8 forks source link

[dattri.benchmark] refactor the benchmark code for mnist #56

Closed TheaperDeng closed 4 months ago

TheaperDeng commented 4 months ago

Description

1. Motivation and Context

Under the benchmark folder, we will have multiple datasets, with some datasets sharing some models (e.g., resnet or GPT). We are thinking about having a benchmark/models folder for the model code and a benchmark/ folder for the training and eval code for each dataset.

2. Summary of the change

  1. Move the Logistic Regression model definition to dattri.benchmark.models.logistic_regression
  2. Move the mnist related functions to dattri.benchmark.mnist
  3. Add a new function create_mnist_dataset to help creating unified mnist dataset.

3. What tests have been added/updated for the change?

tingwl0122 commented 4 months ago

I will also have a look later.

xingjian-zhang commented 4 months ago

Thanks. I will have a look tonight. On May 3, 2024 at 1:14 PM +0800, tingwl0122 @.***>, wrote:

I will also have a look later. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because your review was requested.Message ID: @.***>