Releases: SPOClab-ca/dn3
Releases · SPOClab-ca/dn3
v0.2-alpha
This is associated with the first version of the preprint (i.e. version upon submission for initial publication).
Attached with the release are the pre-trained T-Vectos weights used.
Initial minimal release
This release establishes the major modules and intent of the DN3 package.
The major features include:
- Clearly demarcated contextual representations of BCI datasets
- Dataset -> Thinker -> Recording
- From the former two, an interface for somewhat-standardized data segmentation along informative (e.g. thinker) boundaries (LO/MSO cross validation)
- Dataset -> Thinker -> Recording
- An automatic data loading pipeline, configured using very simple YAML syntax
- API for lazy transformation system
- Limited API for simplified training of DNNs using the above frameworks