fastai

Version: 1.2.2

Summary of issue
Including a @dataclass decorated class breaks getsource in show_doc

Minimal example

• Clone nbdev_tutorial
• Add new cell in 00_core.ipynb

#export
from dataclasses import dataclass

@dataclass
class Foo:
    a: int
    b: str
    c: float

• Run nbdev_build_lib
• Run nbdev_build_docs

nbdev_build_docs

At the moment, it's possible to do

toc: true

to autogenerate a table of contents. I think it would be nice if there were also styling options, e.g. to have that TOC float on one side of the page.

Is there a convenient way to export models to ONNX?

I have a pretty satisfying mmdet.retinanet model with backbone resnet50_fpn_1x but I have a hard time exporting it to ONNX format

I've tried with the pytorch2onnx.py script from mmdetection but it was unsuccessful

and torch.onnx.export doesn't work:

torch.onnx.export(torch_model, 
                  (img, img_metas), 
            

Design a logo for LibreASR and share it here.

To make an open source project cool, it should have a logo 😄

With #151, FastAI.jl is getting high-level interfaces for searching datasets (finddatasets) and loading datasets into task-specific data containers (loaddataset). There is also a new DatasetRecipe that encapsulates configuration for loading a data container and the block information from a path. These recipes can be registered with a dataset so that they can be found using the above high-lev