ignite
ignite copied to clipboard
High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.
## 🚀 Feature When we run baseline configuration : ``` python -m torch.distributed.launch --nproc 2 --use_env -m py_config_runner ./code/scripts/training.py ./configs/train/baseline_resnet101.py ``` When model is instantiated, it downloads ImageNet pretrained weights...
## 🐛 Bug description Below code shows the error: ```python from ignite.engine import Engine, Events from ignite.utils import setup_logger stop_iter = 50 epoch_length = 100 max_epochs = 2 trainer =...
## ❓ Questions/Help/Support Hi, I'm working with a model that increases in complexity during training. To avoid memory issues, I reduce the batch size accordingly at each epoch. This means...
## 🚀 Feature I'd like to discuss a bit about your PR on IterationMetric (https://github.com/Project-MONAI/MONAI/pull/1488). As far as I understand, the idea is to compute metric value on each iteration...
## 🚀 Feature Higher API for training will improve usability of Ignite as Keras did. My implementation [here](https://github.com/Minyus/pipelinex/blob/master/src/pipelinex/ops/ignite_ops.py) could be used as the base code.
## ❓ Questions/Help/Support Hi @vfdev-5 , I am developing distributed evaluation feature, and facing a problem that the `preds` and `labels` on different GPUs don't have the same length, then...
## ❓ Questions/Help/Support Hi @vfdev-5 , I am writing an ignite handler to write the segmentation metrics of every image into 1 CSV file as the summary, for example: ```...
Description: Will enable the following: ```python events = Events.STARTED(args.start_validation) | Events.COMPLETED | Events.EPOCH_COMPLETED(every=3) trainer = ... @trainer.on(events) def run_validation(trainer): # do something ``` To simplify this implementation in a real...
It would be a good idea to add a mAP metric which is often used in object detection. For this the newly added [IoU](https://github.com/pytorch/vision/blob/master/torchvision/ops/boxes.py#L102-L124) function from torchvision could be used.
## 🐛 Bug description Thanks to PRs on mypy annotation by @gruebel we discovered few classes where we have to redesign methods in order to keep consistency with their base...