Max Morrison

[WIP] Decoding from the categorical posterior distribution

7

Implements the proposed alternative to posterior decoding of the network output as suggested in #37. Remaining tasks: - [ ] Test on original data. My tests have all been conducted...

DeepClustering incompatible with PyTorch on Windows

6

#### Description Given PyTorch's lack of support for Python 2 on Windows systems, the DeepClustering module—or any module depending on PyTorch—cannot be imported alongside nussl. #### Versions Windows-10-10.0.17134 ('Python', '2.7.15...

Default rounding induces length mismatch

1

While rounding may be useful for some applications, it harms others. Consider audio at 22050 Hz sampling rate with length 74751 samples. With a hopsize of 256 samples and padding...

#3 demonstrates that stereo audio throws an error. The solution is to convert to mono. A more descriptive error or warning would help.

Should end when the number of feedback responses reaches maximum--but even number of samples reaching maximum would be better.

Automated attention checks

1

While attention checks can be performed manually, we can make it easier by having users specify a directory of attention checks and specify how many checks per participant should be...

To prevent participants from getting additional information by inspecting HTML

Provide a cost estimate on creation. Should give a min and max cost if a prescreening question has a `correct_answer` defined.