guillaume-be

Hello @MikaelCall , This is a good idea - it would be great if you could contribute your code back to the community! I have a few questions on my...

I understand the motivation of speeding up the loop using Rust rather than a pure Python implementation, but I am unsure this is the most effective way to achieve greater...

Hello, Apologies for the delayed feedback. I am trying to implement a model using custom backward pass operations to reduce memory footprint during training (see https://ai.googleblog.com/2020/01/reformer-efficient-transformer.html and the reversible layers)....

Hello @npatsakula , Thank you for working on adding support for NLLB - this will be a very useful addition to this library. I suggest using the distilled verion (600M)...

Hello @ChrisBeeson , The transformer model itself is probably rather straightforward to implement as it seems to use standard Torch operations throughout that are supported in [tch-rs](https://github.com/LaurentMazare/tch-rs). The tricky part...

Hello @Miezhiko , Thank you for the suggestion - I will look into it in the next few weeks. I will probably target the smaller models for an initial implementation...

Hello @npatsakula , You are right I got confused into assuming that `Path` was a subset of `&str`, it does make sense to keep a richer interface. As the library...

> Hello @guillaume-be! I reverted to `AsRef` (doc-comments too). Thanks a lot! looks good to merge - sorry again for the unneeded iterations

@npatsakula I have created a PR to add support for special token maps across the entire collection of tokenizers and vocab to handle both this PR and the ideas suggested...

Hello @yijunyu , Thank you for your interest in the library and using CodeBERT. The RoBERTa architecture is already available in the library -- I have converted the weights and...