Matteo Guarrera comments

Results 9 comments of


                                            Matteo Guarrera

Use the discriminator of the model after training

Thanks @fealho for the answer. I can do that and request a pull if you are interested.

Query about the reproducibility of the Motion Capture dataset in "Scalable Gradients..." (Li et al., 2020)

Is anyone willing to share their implementation to reproduce the results of the paper? :pleading_face: :pray:

the preprocessed data

Dear authors, I would like to access to the pre-processed data for research purposes. email [email protected]

env yaml missing tabulate

we should use the same requirements between yaml and requirements.txt I noticed also that jax 0.4.20 raise some deprecation error. So I would suggest to push everything to jax 0.4.26....

env yaml missing tabulate

I actually create a conda environment using yaml on NERSC, so actually there is no need to keep `requirements.txt`. You can just `conda env create -f env.yml` or `conda env...

Resolution requirement for mixed-precision FNO

I am looking as well for the code to reproduce the experiments of [Guaranteed Approximation Bounds for Mixed-Precision Neural Operators](https://arxiv.org/pdf/2307.15034) I guess @crwhite14, @JeanKossaifi, @rtu715 are the main authors. Thank...

Extremely slow execution of CPU JAX on cluster vs. laptop (reraise_with_filtered_traceback?)

Same here.

Imports in the examples broken

You are right, I guess the error is in the release. The repo is ok, but if you pip install the package it throws the error

Compatibility Bugs

minor bug in the visualization function of `latent_sde_lorentz.py`