Robert Knight
Robert Knight
After experimentation I found that a much more effective approach to fast iteration over non-contiguous dynamic-rank tensors is to iterate over static-rank views of the inner N dimensions and use...
https://github.com/robertknight/rten/pull/713 implemented the lessons from https://github.com/robertknight/rten/issues/11#issuecomment-2611964200 to achieve the main goal. The original idea in this issue, to coalesce dimensions, could still be added in future.
From the current error messages, it is useful to be able to see what authority (and therefore which service or customer) the user accounts belong to. One way that we...
This is currently failing with a few typecheck issues related to vinyl types: ``` node_modules/@types/vinyl/index.d.ts(335,15): error TS2430: Interface 'BufferFile' incorrectly extends interface 'File'. Types of property 'isBuffer' are incompatible. Type...
There is a WASM build, but you have to build from source as it isn't published to eg. `npm` yet. To try this out locally: 1. Clone repository and run...
> Should I leave this open till it's available on npm? Yes, I think so. There isn't an existing tracking issue for publishing an npm package.
Hello @hbmartin, this library reads models in [`.rten` format](https://github.com/robertknight/rten/blob/main/docs/rten-file-format.md) rather than `.onnx`. RTen models are produced from ONNX models using [rten-convert](https://pypi.org/project/rten-convert/). You can find links to the current models in...
SplitToSequence does allow zeros in the `split` input. The documentation later says: > Length of each output. It can be either a scalar(tensor of empty shape), or a 1-D tensor....
This warning is quite normal in PyTorch codebases that haven't been updated very recently, as it was only added in PyTorch v2.4.0. Adding `weights_only=True` to the `torch.load` command should resolve...
Did you modify the alphabet used for classification (`DEFAULT_ALPHABET`)? If so can you post the alphabet you used. I notice that this model has a last output dimension of size...