ignite
ignite copied to clipboard
pytorch
High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.
## 🐛 Bug description Hi @vfdev-5 , We upgraded ignite from 0.4.8 to 0.4.9 in MONAI 0.9.1 recently: https://github.com/Project-MONAI/MONAI/pull/4605. Then got the issue report from user: something changed related to...
Fixes #2623 Description: Check list: - [x] New tests are added (if a new feature is added) - [ ] New doc strings: description and/or example code are in RST...
Related #2631 Description: This is the implementation [WIP] for the NDCG metric. Check list: - [X] Create a basic implementation. - [x] Handle rating ties. - [x] Write and parameterize...
## 🚀 Feature To add some more metrics used in Recommendation Systems with tests that can directly be used. Here is the list (@alykhantejani @vfdev-5 feel free to add more):...
Is your feature request related to a problem? Please describe. When checkpoint saving occurs there should be a check if the object that state_dict() is called on is a ZeroRedundancyOptimizer...
Related to #2466 Description: - Ported gpu hvd tests from circleci to GHA
## Improve tests using pytest parametrize The idea is to improve our testing code using [`pytest.mark.parametrize`](https://docs.pytest.org/en/7.1.x/reference/reference.html#pytest-mark-parametrize). Currently some tests are implemented as : ```python def test_something(): def _test(a): # internal...
## 🚀 Feature Taken from [this colab](https://colab.research.google.com/drive/17Xo80swTxcaZsZOKS08GI-LHYHkg5kud) by @EricZimmermann ### Description #### Problem For large tensors, computing AUC metrics over multiple thresholds is exhaustive and slow. For a sufficiently large...
Fixes #2314 Description: As described in the issue, added a generalized version of EarlyStopping where the `stop_function` and `pass_function` can be user-defined. @sdesrozis let me know if the approach is...
## 🚀 Feature To add a 'no-improvement' handler. It's similar to early stopping, except that early stopping calls the training to terminate, 'no-improvement' handler would've calls a function that user...
