Skip to content
This repository has been archived by the owner on Jun 2, 2020. It is now read-only.

Issues: toodef/neural-pipeline

28 Open 3 Closed
28 Open 3 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Metrics computation is wrong? wontfix This will not be worked on
#33 opened Aug 4, 2019 by elch10
@toodef
1
Add DataProducerPredictor
#32 opened Jun 4, 2019 by toodef
Let user choose a device by himself
#31 opened Jun 3, 2019 by toodef
Improve metrics
#30 opened Jun 3, 2019 by toodef
Add the ability to select a subpart from dataset good first issue Good for newcomers
#29 opened May 13, 2019 by toodef
Extract all console output to Console Monitor feature All features, that will be nicely improve Neural Pipeline good first issue Good for newcomers
#28 opened May 7, 2019 by toodef
StandartStage doesn't exists in train_config/__init__.py bug Something isn't working
#26 opened Apr 9, 2019 by toodef
@toodef
Add Comparable Trainer
#25 opened Apr 1, 2019 by toodef
@toodef
1
Disappearance of metrics_log when resuming bug Something isn't working
#24 opened Mar 29, 2019 by Balykovsky
@toodef
TensorboardMonitor is_continue flag bug Something isn't working
#23 opened Feb 18, 2019 by Balykovsky
@toodef
Add continue from best checkpoint to Predictor
#21 opened Feb 15, 2019 by toodef
Add a guide about execution performance docs All issues, that related to documentation
#20 opened Feb 15, 2019 by toodef
1
Add trainer with asynchronous metrics processing
#18 opened Feb 14, 2019 by toodef
Add system monitor
#17 opened Feb 14, 2019 by toodef
Add turbojpeg for images fast loading
#16 opened Feb 14, 2019 by toodef
Add built-in metrics for classification and segmentation tasks
#15 opened Feb 14, 2019 by toodef
Add FP16 support
#14 opened Feb 14, 2019 by toodef
Add models constructor
#13 opened Feb 14, 2019 by toodef
Add TTA to Predictor good first issue Good for newcomers
#12 opened Feb 14, 2019 by toodef
Add jit training
#11 opened Feb 14, 2019 by toodef
Profile code and optimize as possible
#10 opened Feb 14, 2019 by toodef
Add batch gradients accumulating (for increase batch without increasing memory usage) good first issue Good for newcomers
#9 opened Feb 14, 2019 by toodef
Add best practices guide to documentation docs All issues, that related to documentation
#7 opened Feb 14, 2019 by toodef
Think about parallel losses computation. May be needed when losses computed on CPU.
#6 opened Feb 14, 2019 by toodef
Add delayed losses computation for increase batch size without memory increasing.
#5 opened Feb 14, 2019 by toodef
ProTip! no:milestone will show everything without a milestone.