Thomas Walther

+1

@timfraedrich I just came here because of the same issue - being able to cache offerings and eligibility would be great. Afaics this would require making `RevenueCat.Package` conform to `Codable`.

@NachoSoto afaics one can encode at least an SK2Product using `jsonRepresentation`, but I can't see how to decode it again. I think in general though, being able to cache as...

I've also just run into an issue reading opus files. It seems that reading the last block fails, but the first blocks work fine. Here's an audio file where reading...

@duggelz as a year has past since your last comment, what is your opinion today on how far out this is on the roadmap?

@TobyRoseman Here's a small self-contained script to reproduce it. torch 1.13, coremltools 6.2. ```python import torch import coremltools as ct @torch.jit.script def _padded_size(x): """Pytorch cannot trace x.shape[] code. We therefore...

I actually managed to find a workaround: if I replace `.to(torch.float)` with `.type_as(x)`, it works fine :).

I could reproduce this with a small STFT layer. Keras model: [keras_stft.keras.zip](https://github.com/tensorflow/tensorflow/files/13843255/keras_stft.keras.zip) SavedModel: [keras_stft_tf.zip](https://github.com/tensorflow/tensorflow/files/13843261/keras_stft_tf.zip) I don't have the TF code to create said STFT, as it was converted from Torch...

Another option is to create a second repository for this, like a mirror. I've found https://github.com/malcommac/sentry-cocoa-sdk-xcframeworks/ that does exactly that, but trusting a binary framework from an unofficial source is...