Documentation Notes

[JLenzy]

• [ ] Possible to have a --requirements.txt file?
• [ ] (3.1) minor correction, on my system I had to go up 2 directories to get it to work:

gmd_pickle_path="../../data/gmd/resources/storedDicts/groove_2bar-midionly.bz2pickle"))

• [ ] 3.1.3 I was slightly confused by the 'max_len' argument; wasn't sure if this was somehow related to batch size or the sequence length itself. Maybe clarifying this + adding a brief comment about the padding/slicing it does if len != maxlen

Will update this as I go along..

[JLenzy]

Section 2: Models

• [ ] Question on embedding size: Once this has been set, does it remain the same despite the size of the input vector? I am thinking in regards to NLP, where one sentence might be 6 words, and the next could be 13. If embed_size = 32 for example, would this stay static throughout any type of inference length? Does it need to be larger than the largest possible input matrix?