neurostatslab vocalocator issues

neurostatslab / vocalocator

Deep neural networks for sound source localization and vocalization attribution.

MIT License

2 stars 0 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Implement Model Output types for training on 3d locations

#68 Aramist opened 5 months ago
1
See if the new numpy release breaks anything

#67 Aramist closed 5 months ago
0
Create jupyter notebook for advanced usage and visualization on video

#66 Aramist opened 5 months ago
2
Prettify the Installation section of the readme

#65 Aramist closed 5 months ago
0
Add section on config fields to readme

#64 Aramist opened 5 months ago
0
Add a quick-start section to readme

#63 Aramist closed 5 months ago
1
Add install instructions for multiple package managers

#62 Aramist closed 5 months ago
0
Change project name to Vocalocator

#61 Aramist closed 5 months ago
0
Combine RIR dataset and VocalizationDataset into one

#60 Aramist closed 6 months ago
0
Save test_set indices in addition to train and val when initializing dataloaders

#59 Aramist closed 6 months ago
0
Add support for 3d locations

#58 Aramist opened 6 months ago
0
Add support for 3d locations

#57 Aramist closed 6 months ago
0
Remove inconsistencies in the main branch

#56 Aramist closed 5 months ago
2
Cleanup for official release

#55 Aramist closed 10 months ago
0
Final changes for first official release

#54 Aramist closed 10 months ago
1
Enable models to output points outside arena grid

#53 amanchoudhri closed 5 months ago
0
Running eval generates a model directory where none is necessary

#52 Aramist closed 5 months ago
0
Use larger grid for calibration calculation

#51 amanchoudhri opened 1 year ago
0
Calculate calibration using a larger grid than arena size

#50 amanchoudhri closed 5 months ago
0
DRAFT: Add class to package together multiple datasets

#49 amanchoudhri opened 1 year ago
0
Add ability to train on multiple `GerbilVocalizationDataset` objects at once

#48 amanchoudhri closed 5 months ago
0
Enable `MDNOutput` to return pmf mode instead of weighted average of individual point estimates

#47 amanchoudhri closed 5 months ago
0
Remove erroneous error catching on label outside arena dims

#46 amanchoudhri closed 5 months ago
0
Use weighted average in `MDNOutput._point_estimate` instead of simple average

#45 amanchoudhri closed 1 year ago
0
Use weighted average for MDN point estimation

#44 amanchoudhri closed 1 year ago
0
Add temp scaling to `ProbabilisticOutput`

#43 amanchoudhri closed 1 year ago
0
Separate loss function specification from `build_model`

#42 amanchoudhri closed 5 months ago
0
Avoid edge case where Trainer wraps `ReduceLROnPlateau` in `SequentialLR`

#41 amanchoudhri opened 1 year ago
0
`ReduceLROnPlateau` scheduler cannot be wrapped in `SequentialLR`

#40 amanchoudhri closed 5 months ago
0
Add `load_weights` methods to `GerbilizerArchitecture` and subclasses

#39 amanchoudhri closed 1 year ago
2
Retry 2d confidence map output

#38 Aramist closed 5 months ago
0
Fix save directory bug

#37 Aramist closed 1 year ago
0
Add nicer plot outputs during training and assessment

#36 amanchoudhri closed 5 months ago
0
Require crop length parameter

#35 Aramist closed 1 year ago
1
Remove variable length input option

#34 Aramist closed 1 year ago
0
Appropriately reweight loss function based on arena size?

#33 amanchoudhri closed 5 months ago
0
Add option to Trainer to save best CALIBRATED model

#32 amanchoudhri closed 5 months ago
0
Add temperature scaling parameter to ProbabilisticOutput

#31 amanchoudhri closed 1 year ago
0
Enable specifying LR Schedule in model config

#30 amanchoudhri closed 1 year ago
0
Create unified way to load model weights

#29 amanchoudhri closed 1 year ago
1
Modernize Densenet implementation

#28 Aramist closed 1 year ago
0
Allow user to specify learning rate schedule in model config

#27 amanchoudhri closed 1 year ago
0
Add learning rate warmup

#26 Aramist closed 1 year ago
0
Have separate IterableDataloder and Dataloader for variable-length and fixed-length data

#25 Aramist closed 1 year ago
1
Create a more cohesive training/evaluation interface

#24 amanchoudhri closed 5 months ago
0
Create utility classes for model output

#23 amanchoudhri closed 1 year ago
2
Reimplement data-augmentations

#22 Aramist closed 1 year ago
6
`run_eval` function in `main.py` brittle to output shape

#21 amanchoudhri closed 1 year ago
0
Rehaul data augmentation pipeline

#20 Aramist closed 1 year ago
0
Add units to output arrays

#19 Aramist closed 1 year ago
1