Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / CarloLucibello/Tsunami.jl issues and pull requests
#66 - run GC every epoch
Issue -
State: open - Opened by CarloLucibello 6 months ago
#65 - relax requirement that FluxModule is mutable
Issue -
State: open - Opened by CarloLucibello 6 months ago
#64 - add VAE example
Pull Request -
State: closed - Opened by CarloLucibello 6 months ago
#63 - stop training on Ctrl+C
Issue -
State: open - Opened by CarloLucibello 6 months ago
#62 - follow `@functor` -> `@layer` changes in Flux
Pull Request -
State: closed - Opened by CarloLucibello 7 months ago
- 1 comment
#61 - use Flux.get_device
Pull Request -
State: closed - Opened by CarloLucibello about 1 year ago
#60 - improve tests for Metal and add Foil example
Pull Request -
State: closed - Opened by CarloLucibello about 1 year ago
#59 - Flux0.14 support + AMDGPU and Metal GPU support
Pull Request -
State: closed - Opened by CarloLucibello about 1 year ago
- 1 comment
#58 - deprecation for fit(ckpt_path, trainer, ...)
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#57 - accumgrad and clipgrad
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
#56 - save Flux.state(model) in checkpoints
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#55 - save the model_state in the checkpoint instead of the full model
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#54 - make Tsunami.fit the main API
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#53 - deprecate `fit!` in favor of `fit`
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#52 - Add Tsunami.fit + foreach_trainable
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#51 - simplify show and fix functor FluxModule
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#50 - default gpu device should be the one with largest free memory
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#49 - add a get_device method
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#48 - bert finetuning example
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#47 - more hooks + precision setting + other stuff
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
#46 - use MLDatasets in the README example
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#45 - add a NLP example
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#44 - introduce the accelerator Foil
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
#43 - create a lightweight training framework
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#42 - on_before_pullback_call
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#41 - add validate and on_start_*_epoch hooks
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#40 - implement Tsunami.test
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#39 - add docs for gradient clipping
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 2 comments
#38 - add accumulate gradient
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#37 - add validate and test methods
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#36 - overwrite progbar at each epoch
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#35 - add more helpful errore message for not implemented methods
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#34 - TagBot trigger issue
Issue -
State: closed - Opened by JuliaTagBot over 1 year ago
- 8 comments
#33 - support multi-GPU training
Issue -
State: open - Opened by zsz00 over 1 year ago
- 2 comments
#32 - reuse the progress bar, don't create a new one for each epoch
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#31 - support scheduler configuration
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#30 - schedulers should become lr_schedulers?
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#29 - change `training_*` to `train_*` etc..
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#28 - check licensing issues with pytorch lightning
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#27 - make batch_idx optional in training_step etc...
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#26 - Tsunami.seed!
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
#25 - add DataModule
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#24 - [logging] weight with batchsize on_epoch reduction
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#23 - change progress bar color to white and blue
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#22 - account for batch_size when logging on_epoch
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#21 - create logging infrastructure
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#20 - audit API
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#19 - deprecate training_epoch_end
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#18 - check and improve documentation
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#17 - add other loggers
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#16 - tag a first release
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#15 - add seed_everything
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#14 - add graph neural networks example
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#13 - add log function
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
Labels: priority
#12 - replace bson checkpointing with jld2?
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#11 - describe in pseudocode the `fit!` method
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#10 - Callbacks support
Pull Request -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#9 - no training progress bar is shown for small number of batches
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#8 - define show method for generic FluxModel
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#7 - make fit_state its own type
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
#6 - add callbacks support
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#5 - cleaner validation printing
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#4 - mixed precision training support
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#3 - change the name of the library?
Issue -
State: closed - Opened by CarloLucibello over 1 year ago
- 1 comment
#2 - missing arguments for Trainer compared to pytorch-lightning
Issue -
State: open - Opened by CarloLucibello over 1 year ago
#1 - cannot checkpoint models containing closures
Issue -
State: open - Opened by CarloLucibello over 1 year ago