Richard Harris

The Sup v5 models use the new transformer architecture which is slower but more accurate than existing models. There will be a significant increase in performance as we continue to...

Hi @wtfeng111, The modbase tag specification annotations state the strandedness of the modifications. This information allows for modifications on 2d reads to be annotated by indicating on which strand the...

The model used by dorado correct is the same as the one use by HERRO for r10.4 data. As such dorado correct is not supported for r9.4.1 data unfortunately. We're...

At this time dorado correct only has support for the R10.4.1 herro model. Kind regards, Rich

Hi @changsarahl, At this time we're not planning on adding support for the R9.4.1 condition / models in `dorado correct` Kind regards, Rich

Hi @sivico26, [Dorado-0.8.0](https://github.com/nanoporetech/dorado/releases/tag/v0.8.0) is now released with this additional functionality split the alignment and inference stages. There's also the new resume functionality too! Thanks for the feature request! Kind regards,...

Hi @wshropshire, We'll discuss adding this as a feature. Similar issues to add more context: https://github.com/nanoporetech/dorado/issues/1077 https://github.com/nanoporetech/dorado/issues/1026 Kind regards, Rich

@fen2323, It's possible although not recommended as dorado correct has very high resource requirements. [Dorado-0.8.0](https://github.com/nanoporetech/dorado/releases/tag/v0.8.0) allows the splitting of the CPU (alignment) and GPU (inference) stages if that's what your...

Dorado 0.8.0 contains improvements to the auto batch size calculation and how memory is allocated when using modbase models. Further improvements will be made as we're continually working on the...