I want to use FasttextEmbedder. How should I guess that it requires installing a separate requirements file? By the way, installing a requirements file from git without cloning is not that obvious. It should be explicitly mentioned in the documentation. Another problem is the lack of separate fasttext embedder config for sentences, in contrast to elmo embedder.

For example, documentation of AudioToSpectrogramPreprocessor is missing.

We investigated that with @redoctopus and she has found that nemo_asr.rst doesn't include several modules.

https://github.com/NVIDIA/NeMo/blob/master/docs/sources/source/collections/nemo_asr.rst

So the bigger issue includes: taking a look at all implemented components and checking whether the associated rst files are

not a bug report per se

I'm wondering how spacy/chinese models compares with the stanza project?
Stanza already provides chinese support with many features
https://stanfordnlp.github.io/stanza/models.html

that has a chinese (simplified) model and provides dep-parser, lemma and other basic NLP features.

I'm a bit confused as it uses spacy for tokenization:
https://stanfordnlp.github.io/

Create a markdown document with some explanations on how level generation works in BabyAI and how to create custom levels. This would go under /docs.

More broadly, I'd like to split up the README into more sub-documents, as I've done in the MiniWorld repo. Keep only introduction, installation and basic usage into the main README