Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / mlverse/luz issues and pull requests

#145 - How to use luz with custom optimizer?

Issue - State: open - Opened by icejean about 2 months ago

#144 - cpp_tensor_from_buffer error

Issue - State: open - Opened by maxwell-geospatial about 2 months ago

#143 - Apparent issue with macOS Sonoma 14.4.1

Issue - State: open - Opened by DenaJGibbon 2 months ago - 1 comment

#142 - rerun the same 'setup' codes a second time results in error

Issue - State: closed - Opened by Backlonhw2468 3 months ago - 5 comments

#141 - MacOs fixes

Pull Request - State: closed - Opened by dfalbel 4 months ago

#140 - Add scheduler reduce on plateau

Pull Request - State: closed - Opened by SvenVw 7 months ago - 1 comment

#139 - Generative

Pull Request - State: closed - Opened by dfalbel 9 months ago

#138 - add chargpt model example

Pull Request - State: closed - Opened by dfalbel 10 months ago

#137 - Lr scheduler state

Pull Request - State: closed - Opened by dfalbel 10 months ago

#136 - luz_callback_auto_resume doesn't resume ctx$epoch

Issue - State: closed - Opened by pradosj 10 months ago - 4 comments

#135 - Add support for iterable datasets.

Pull Request - State: closed - Opened by dfalbel 10 months ago

#134 - Add text classification example

Pull Request - State: closed - Opened by dfalbel 10 months ago

#133 - CRAN and GitHub versions failed to install WIN11

Issue - State: open - Opened by oguchihy 11 months ago - 1 comment

#132 - Fix "mmetrics" typo

Pull Request - State: closed - Opened by jonthegeek 11 months ago - 1 comment

#130 - `predict`: How to get embedding layer?

Issue - State: open - Opened by talegari about 1 year ago - 3 comments

#129 - Mention how to use metrics in the docs

Pull Request - State: closed - Opened by dfalbel about 1 year ago

#128 - luz metrics do not work with optim_lbfgs.

Issue - State: closed - Opened by wbzyl about 1 year ago - 2 comments

#127 - Add callback to use automatic mixed precision.

Pull Request - State: closed - Opened by dfalbel about 1 year ago

#126 - Release luz 0.4.0

Issue - State: closed - Opened by dfalbel about 1 year ago

#125 - Reduce profile overhead

Pull Request - State: closed - Opened by dfalbel about 1 year ago

#124 - Faster loss averaging

Pull Request - State: closed - Opened by dfalbel about 1 year ago

#123 - Evaluate different metrics

Pull Request - State: closed - Opened by dfalbel about 1 year ago

#122 - Improve garbage collection opportunities for datalaoder workers

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#121 - Bind context to the `.__enclos_env__`.

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#120 - Handle deprecations.

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#119 - always respect `cpu` flag

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#118 - Add tfevents integration

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#117 - Drop last

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#116 - Remove `coro::generator()` call

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#115 - Device method

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#114 - Context aliases

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#113 - How to just continue training

Issue - State: closed - Opened by GitHubGeniusOverlord over 1 year ago - 2 comments

#112 - Specialized metrics

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#111 - MixUp callback should be configurable

Issue - State: closed - Opened by dfalbel over 1 year ago

#110 - Avoid computing metrics on train data

Issue - State: closed - Opened by dfalbel over 1 year ago

#109 - Debug errors in callbacks/metrics

Issue - State: closed - Opened by dfalbel over 1 year ago

#108 - Examples

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#107 - Autoresume

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#106 - Update luz examples

Issue - State: closed - Opened by dfalbel over 1 year ago

#105 - Add documentation for prediction and evaluation callbacks.

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#104 - M1 mac support

Pull Request - State: closed - Opened by dfalbel over 1 year ago

#103 - predict.luz_module_fitted callbacks help

Issue - State: closed - Opened by jonthegeek almost 2 years ago

#102 - Rstudio crashes when fitting using luz

Issue - State: closed - Opened by MislavSag almost 2 years ago - 8 comments

#101 - MNIST example error: missing value where TRUE/FALSE needed

Issue - State: closed - Opened by MislavSag almost 2 years ago - 1 comment

#100 - Release luz 0.3.1

Issue - State: closed - Opened by dfalbel almost 2 years ago

#99 - Release luz 0.3.0

Issue - State: closed - Opened by dfalbel almost 2 years ago

#98 - progress bar for `lr_finder`.

Pull Request - State: closed - Opened by dfalbel about 2 years ago

#97 - Lr finder improvements

Pull Request - State: closed - Opened by dfalbel about 2 years ago - 1 comment

#96 - Fix typos in fit help.

Pull Request - State: closed - Opened by jonthegeek about 2 years ago - 1 comment

#95 - Add callback to use the best model

Pull Request - State: closed - Opened by dfalbel about 2 years ago

#94 - Fix README typos.

Pull Request - State: closed - Opened by jonthegeek about 2 years ago - 3 comments

#93 - Add backward argument to `setup`.

Pull Request - State: closed - Opened by dfalbel about 2 years ago

#92 - Weights

Issue - State: open - Opened by dfalbel about 2 years ago - 1 comment

#91 - FIgure out how to modify how `backward()` is called.

Issue - State: closed - Opened by dfalbel about 2 years ago - 1 comment

#90 - Add callback for gradient clipping inspired by FastAI's.

Pull Request - State: closed - Opened by dfalbel over 2 years ago

#89 - Learning rates tested by lr_finder() run into saturation

Issue - State: closed - Opened by EmKarr over 2 years ago - 1 comment

#88 - Stop/resume model fitting

Issue - State: closed - Opened by pradosj over 2 years ago

#87 - lr_finder changes minibatch size

Issue - State: closed - Opened by jtrecenti over 2 years ago - 1 comment

#86 - Examples dont run on their own

Issue - State: closed - Opened by GitHubGeniusOverlord over 2 years ago - 1 comment

#85 - Update learning rate exponentially and maybe, add smoothing to loss plot

Pull Request - State: closed - Opened by skeydan over 2 years ago - 2 comments

#84 - move weight reshaping inside nnf_mixup

Pull Request - State: closed - Opened by skeydan over 2 years ago

#83 - Consider only computing metrics for the validation set.

Issue - State: closed - Opened by dfalbel over 2 years ago

#82 - add mixup callback and mixup cross entropy loss

Pull Request - State: closed - Opened by skeydan over 2 years ago - 9 comments

#81 - Clean up wording in get-started vignette.

Pull Request - State: closed - Opened by jonthegeek over 2 years ago - 1 comment

#80 - add label smoothing

Pull Request - State: closed - Opened by skeydan over 2 years ago - 1 comment

#79 - return the mod instead due to testthat changes.

Pull Request - State: closed - Opened by dfalbel over 2 years ago

#78 - Error with last batch that is smaller than the others

Issue - State: closed - Opened by skeydan almost 3 years ago - 4 comments

#77 - Fix add = TRUE order in `on.exit`

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#76 - Bugfix/early stopping

Pull Request - State: closed - Opened by dfalbel almost 3 years ago - 2 comments

#75 - Move `master` branch to `main`

Issue - State: closed - Opened by jennybc almost 3 years ago - 1 comment

#74 - Leak the `fit` function environment

Issue - State: closed - Opened by dfalbel almost 3 years ago - 8 comments

#73 - Implements `evaluate`

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#72 - Implement `evaluate` so we can quickly compute metrics in a specific dataset.

Issue - State: closed - Opened by dfalbel almost 3 years ago - 1 comment

#71 - Feature/dataloader options

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#70 - Training data should be shuffled by default

Issue - State: closed - Opened by dfalbel almost 3 years ago

#69 - support for proportion of valid_data + documentation improvements.

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#68 - Support for `valid_data = p`

Issue - State: closed - Opened by dfalbel almost 3 years ago

#67 - Feature/numerics

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#66 - Adds `as_dataloader` to cast data into dataloader

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#65 - Unet example

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#63 - Early stopping with patience=1 shouldn't stop in the first epoch

Issue - State: closed - Opened by dfalbel almost 3 years ago

#62 - Release luz 0.2.0

Issue - State: closed - Opened by dfalbel almost 3 years ago

#61 - set old seed behavior to make tests pass with newer torch version

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#60 - round batch size usign floor() and change order of start and end lr

Pull Request - State: closed - Opened by cmcmaster1 almost 3 years ago - 1 comment

#59 - add lr_finder

Pull Request - State: closed - Opened by cmcmaster1 almost 3 years ago - 2 comments

#58 - support for choosing CUDA device index if multiple GPUs present

Pull Request - State: closed - Opened by cmcmaster1 almost 3 years ago

#57 - Initial work on tracing callback

Pull Request - State: open - Opened by dfalbel almost 3 years ago

#56 - Implement luz_callback_validation_check

Pull Request - State: open - Opened by mattwarkentin almost 3 years ago

#55 - Provide support for minimum and maximum number of epochs

Pull Request - State: closed - Opened by mattwarkentin almost 3 years ago - 1 comment

#54 - Fixes progress bars for predictions.

Pull Request - State: closed - Opened by dfalbel almost 3 years ago

#53 - Minimum and maximum number of epochs

Issue - State: closed - Opened by mattwarkentin almost 3 years ago - 2 comments

#52 - Fixes csv logger and validation progress bar length

Pull Request - State: closed - Opened by mattwarkentin almost 3 years ago - 1 comment

#51 - Issue with length of progress bar for validation loop

Issue - State: closed - Opened by mattwarkentin almost 3 years ago

#49 - Support for mixed precision

Issue - State: open - Opened by dfalbel about 3 years ago

#48 - Release/v0.1.0

Pull Request - State: closed - Opened by dfalbel about 3 years ago

#47 - Release luz 0.1.0

Issue - State: closed - Opened by dfalbel about 3 years ago

#46 - Add a simple plot method for the fitted module

Pull Request - State: closed - Opened by dfalbel about 3 years ago