FluxML/FastAI.jl

Guides for block level customization

yl4070 opened this issue · 1 comments

Hi,
Currently, the customization tutorial is based on low level API, it would be nice to have a similar guides but based on block level interface, covering materials such as how to write customized blocks for different use cases, and what functions need to be implemented for the new block to work properly. It appears to be a big jump from using existing blocks to customize everything.

Personally, I'm struggling to figure out how to make new blocks work properly by looking at the source code, so I'm really looking forward to such guides.

Yeah, documentation for that hasn't been written yet. If you want to implement new Encodings for existing blocks, for now the best resource is the docstring for Encoding. You can also look at and use FastAI.testencoding as a reference for the encoding being implemented properly. If you have any specific questions or stumbling blocks, feel free to ask, though!

I'll leave this open and give an update once a proper guide is in the docs.