Lu Jiang

You can find these features using a more recent tool: https://cleanlab.ai/ On Tue, May 10, 2022 at 12:28 AM nazaretl ***@***.***> wrote: > Hi, > thanks for sharing your implementation....

We first train our model for 18 epochs on the noisy dataset. Then we use the model to evaluate on another small dataset, where we have some clean labels. The...

Details are in https://github.com/google/mentornet/blob/master/TRAINING.md clean label column: ground-truth labels on small clean dataset noisy label column: given labels on the current noisy dataset loss column: loss computed using the noisy...

The authors are supposed to generate their own csv files from their models. On Mon, May 13, 2019, 2:27 AM wffzxyl wrote: > Details are in > https://github.com/google/mentornet/blob/master/TRAINING.md > clean...

Could you please have another try. The ckpt files should be accessible. I was using https://colab.sandbox.google.com/

This experiment only shows in the supplementary and does not belong to the main paper. We'd love to release the code but it may require additional approvals. We will try...

Thank you all for your interest and patience. Each time you ask, I consult with our internal lawyers and seek the necessary approvals. We genuinely hope to release the models...

Same question here. Will you update the README and state we should evaluate on 50889?