tch-rs icon indicating copy to clipboard operation
tch-rs copied to clipboard

Rust bindings for the C++ api of PyTorch.

Results 213 tch-rs issues
Sort by recently updated
recently updated
newest added

The new M1 Mac's (soon to be all of them) have aarch64 rather than X64 and libtorch is an X64 build under Rosetta 2. This means that it's impossible to...

Dear team, On macOS arm64 M1 pro, I have the following error: Compiling tch v0.8.0 (/Users/jianshuzhao/Github/tch-rs) error: linking with `cc` failed: exit status: 1 | = note: "cc" "-arch" "arm64"...

Dear torch rust team, The newest version of PyTorch (https://pytorch.org/blog/introducing-accelerated-pytorch-training-on-mac/) also support apple M1 devices via the MPS interface provided by apple. I am wondering whether this is also doable...

Is there a way to save and load layers so training can be paused and resumed? I am unable to currently find a way with tch-rs.

Hi, I'm trying to run a program using rust-bert crate but unfortunately I'm getting this error. The error appears when I try to import from rust-bert crate, if I don't...

Hi. Thanks for this amazing lib. Just stumbled upon the old issue where default arguments are not exposed. https://github.com/LaurentMazare/tch-rs/issues/132 https://github.com/LaurentMazare/tch-rs/issues/133 https://github.com/LaurentMazare/tch-rs/issues/350 Here the use case is for new the bloom...

I get a model from detectron2 of mask-rcnn . And it is panicked when i load the model by tch::CModule . And the panicked infomation is : > thread 'main'...

Currently it's a bit difficult to create an interface that operates on `CModule`. You may also want to accept `TrainableCModule` since they implement most of the same methods, but there...

1.I have PyTorch 1.10 that works natively on my M1 Mac I've downloaded libtorch from the website 2. I've added tch = "0.6.1" to cargo.toml and when run cargo build...

## Summary - add `create_class_is` for `tch::jit::CModule` - allow creating custom jit object - usage: ```rust // Load the TorchScript model to the target device. let model = tch::CModule::load_data_on_device(&mut model_bytes.as_slice(),...