Super-gradients

New features
predict() support for pose estimation (PR: [1142](https://github.com/Deci-AI/super-gradients/pull/1142)) and classification (PR: [#1220](https://github.com/Deci-AI/super-gradients/pull/1220))
The possibility to log scalar values with explicit time units associated with it: log losses per step, use TimerCallback that measures and log batch time (PR: [1181](https://github.com/Deci-AI/super-gradients/pull/1181))
Support torch compile (PR: [1172](https://github.com/Deci-AI/super-gradients/pull/1172))
Class agnostic NMS (PR: [1232](https://github.com/Deci-AI/super-gradients/pull/1232))
Allow to compute & log per-class mAP scores (PR: [1238](https://github.com/Deci-AI/super-gradients/pull/1238))
Small features
Add batch_size option to predict (PR: [1273](https://github.com/Deci-AI/super-gradients/pull/1273))
Allows to pass user-defined target of IoU thresholds without asking user to edit IoUThresholds enum in SG codebase (PR: [1223](https://github.com/Deci-AI/super-gradients/pull/1223))
Add DetectionVerticalFlip for bounding box detection augmentation (PR: [1237](https://github.com/Deci-AI/super-gradients/pull/1237))
Support the option for multiple ignored indices for segmentation metrics (PR: [1177](https://github.com/Deci-AI/super-gradients/pull/1177))
Simpify torch schedulers integration (PR: [1230](https://github.com/Deci-AI/super-gradients/pull/1230))
support single output in YoloX NMS (PR: [813](https://github.com/Deci-AI/super-gradients/pull/813))
Compute the best threshold for each class in an efficient manner without added loops (PR: [973](https://github.com/Deci-AI/super-gradients/pull/973))
Bug fix
Hotfix/alg 1470 drop boxes padding by (PR: [1107](https://github.com/Deci-AI/super-gradients/pull/1107))
Fix evaluate from recipe by (PR: [1170](https://github.com/Deci-AI/super-gradients/pull/1170))
fixed bug of whitespace by (PR: [1173](https://github.com/Deci-AI/super-gradients/pull/1173))
Fix collections.Iterable -> typing.Iterable to fix crash in python 3.10 by [BloodAxe](https://github.com/BloodAxe) in [#1178](https://github.com/Deci-AI/super-gradients/pull/1178)
Fix missing encoding (PR: [1185](https://github.com/Deci-AI/super-gradients/pull/1185))
default quantization params set in qat from config (PR: [1192](https://github.com/Deci-AI/super-gradients/pull/1192))
Fix mkdirs in checkpoint (PR: [1198](https://github.com/Deci-AI/super-gradients/pull/1198))
Fix DEKR’s replace_head & improve repr for keypoints transforms (PR: [1227](https://github.com/Deci-AI/super-gradients/pull/1227))
Getting rid of “module.” heritage (PR: [1184](https://github.com/Deci-AI/super-gradients/pull/1184))
Fix LayerNorm have a bias parameter attribute but is not instance of torch primitive modules (PR: [1229](https://github.com/Deci-AI/super-gradients/pull/1229))
moved exceptions from training (PR: [1260](https://github.com/Deci-AI/super-gradients/pull/1260))
Added explicit casting of input in predict pipeline. (PR: [1281](https://github.com/Deci-AI/super-gradients/pull/1281))
kd ema model unrwap model fix (PR: [1283](https://github.com/Deci-AI/super-gradients/pull/1283))
New Contributors
[allankouidri](https://github.com/allankouidri) made their first contribution in [#1185](https://github.com/Deci-AI/super-gradients/pull/1185)
[djm93dev](https://github.com/djm93dev) made their first contribution in [#1199](https://github.com/Deci-AI/super-gradients/pull/1199)
[danielafrimi](https://github.com/danielafrimi) made their first contribution in [#1220](https://github.com/Deci-AI/super-gradients/pull/1220)
[RanZilberstein](https://github.com/RanZilberstein) made their first contribution in [#1228](https://github.com/Deci-AI/super-gradients/pull/1228)
[itaylevy-deci](https://github.com/itaylevy-deci) made their first contribution in [#973](https://github.com/Deci-AI/super-gradients/pull/973)
[LukeAI](https://github.com/LukeAI) made their first contribution in [#1237](https://github.com/Deci-AI/super-gradients/pull/1237)
[jorgectf](https://github.com/jorgectf) made their first contribution in [#1240](https://github.com/Deci-AI/super-gradients/pull/1240)
[jacobmarks](https://github.com/jacobmarks) made their first contribution in [#1278](https://github.com/Deci-AI/super-gradients/pull/1278)

What's Changed
* Hotfix/sg 000 fix doc typo rf100 by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/897
* Fixed false-positive warning message by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/886
* Update deprecate -> deprecated by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/906
* Fix not-closed tqdm by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/909
* Update of PoseEstimation model by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/872
* Update the notebooks to install super_gradients==3.1 by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/924
* Fix the bug with YOLONAS of not supporting overriding in_channels by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/923
* Feature/sg 000 no explicit pil by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/908
* Force predictions as list by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/945
* Feature/sg 850 support torch 2.0 by shaydeci in https://github.com/Deci-AI/super-gradients/pull/944
* Bug/sg 764 wrong ckpt when resuming with external ckpt root dir by shaydeci in https://github.com/Deci-AI/super-gradients/pull/948
* Fix typo in welcome.html by eltociear in https://github.com/Deci-AI/super-gradients/pull/941
* Add detection prediction tutorial by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/943
* Update YoloNASQuickstart.md by avideci in https://github.com/Deci-AI/super-gradients/pull/970
* Fix preprocessing by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/972
* Fix train_loader not initalized properly by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/981
* Fix empty mask in MaskAttentionLoss by lkdci in https://github.com/Deci-AI/super-gradients/pull/982
* changed requirement by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1002
* Fix convert_to_onnx to correctly handle case when input_shape is None by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/989
* Feature/sg 757 resume for spots by shaydeci in https://github.com/Deci-AI/super-gradients/pull/870
* Cityscapes AutoLabelling dataset by lkdci in https://github.com/Deci-AI/super-gradients/pull/1000
* Predict on fused model by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/998
* Proposal of issue template by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/1018
* Fix doc by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/1019
* Fixed missing encoding (Issue 999) by T0T4R4 in https://github.com/Deci-AI/super-gradients/pull/1044
* Cache model from platform locally by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1009
* Feature/infra 000 nightly by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1051
* Added CoreML export notebook & method for Apple devices. by avideci in https://github.com/Deci-AI/super-gradients/pull/1068
* Bug/sg 861 decouple qat from train from config by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1001
* updated contributing.md by shaydeci in https://github.com/Deci-AI/super-gradients/pull/965
* Draw most likely bbox on top (1061) by laszlovandenhoek in https://github.com/Deci-AI/super-gradients/pull/1063
* replace documentation 'onnx' with 'coreml' by shanibenbaruch in https://github.com/Deci-AI/super-gradients/pull/1074
* Fix deprecated usage of pycocotools by fcakyon in https://github.com/Deci-AI/super-gradients/pull/1052
* YoloNAS benchmarking tutorial by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1076
* Fix arch_param deprecated by Louis-Dupont in https://github.com/Deci-AI/super-gradients/pull/1084
* fix deprecated naming for stdc by ofrimasad in https://github.com/Deci-AI/super-gradients/pull/1094
* fix: Plot boxes by yurkovak in https://github.com/Deci-AI/super-gradients/pull/1097
* Added link to new QAT fine-tuning notebook (YoloNAS) + removed github notebooks by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1100
* Hotfix/sg 000 fix breaking changes in some imports by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1101
* Hotfix/sg 000 dekr weights hotfix by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1102
* Fixed resuming training with DDP when average_best_models=true by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1111
* Update deci-lab-client usage by roikoren755 in https://github.com/Deci-AI/super-gradients/pull/1047
* Feature/sg 815 fix override dataset params by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1092
* Bug/sg 896 add deprecation for previous breaking changes by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1121
* Actualize pose estimatino zoo. by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1114
* Typo fix in modify_params_for_qat by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1129
* Bug/sg 897 support for min samples with ddp by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1128
* strict_load fixed in recipe default checkpoint_params by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1130
* bump version by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1132
* Update Weights & Biases Integration + Add Prediction Visualization for Object Detection by soumik12345 in https://github.com/Deci-AI/super-gradients/pull/1010
* Bug/sg 000 fix tests by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1131
* Allow checkpoint_path to be an url (Required for transfer learning) by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1138
* Fix incorrect use of collections.OrderedDict in type annotation by BloodAxe in https://github.com/Deci-AI/super-gradients/pull/1141
* Bug/sg 000 merge failure for datasetparams by shaydeci in https://github.com/Deci-AI/super-gradients/pull/1140

New Contributors
* eltociear made their first contribution in https://github.com/Deci-AI/super-gradients/pull/941
* T0T4R4 made their first contribution in https://github.com/Deci-AI/super-gradients/pull/1044
* laszlovandenhoek made their first contribution in https://github.com/Deci-AI/super-gradients/pull/1063
* fcakyon made their first contribution in https://github.com/Deci-AI/super-gradients/pull/1052
* soumik12345 made their first contribution in https://github.com/Deci-AI/super-gradients/pull/1010

**Full Changelog**: https://github.com/Deci-AI/super-gradients/compare/3.1.1...3.1.2

This GitHub Release was done automatically by CircleCI

This GitHub Release was done automatically by CircleCI

Features
- Object Detection (Yolox, PPyoloE) predict function on images, videos, GIF, and folders paths. The predict function works out of the box, no need to define pre and post-processing. Both pre and post-processing are taken from the training. Examples [here](https://github.com/Deci-AI/super-gradients/tree/master/src/super_gradients/examples/predict). PR: [#815](https://github.com/Deci-AI/super-gradients/pull/815), [#804](https://github.com/Deci-AI/super-gradients/pull/804), [#807](https://github.com/Deci-AI/super-gradients/pull/807), [#829](https://github.com/Deci-AI/super-gradients/pull/829), [#845](https://github.com/Deci-AI/super-gradients/pull/845)

- Support yolov5 format detection dataset by for YoloX, PPyoloE https://github.com/Deci-AI/super-gradients/pull/847

- Base recipes for [PPYoloE](https://github.com/Deci-AI/super-gradients/blob/master/src/super_gradients/recipes/roboflow_ppyoloe.yaml) and [Yolox](https://github.com/Deci-AI/super-gradients/blob/master/src/super_gradients/recipes/roboflow_yolox.yaml) of RF100 dataset.

- [DetectionOutputAdapter tutorial](https://github.com/Deci-AI/super-gradients/blob/master/documentation/source/DetectionOutputAdapter.md). The DetectionOutputAdapter is a class that converts the output of a detection model into a user-appropriate format.

- [Segformer model and recipe](https://github.com/Deci-AI/super-gradients/pull/754)
- Predict function on images and videos - taking the pre+post processing from the training recipe.
- [QAT&PTQ notebooks](http://bit.ly/3KrN6an)
- Added student_adapter feature to KDModule, which lets the user pass a manipulated version of the input to the student model. PR: [820](https://github.com/Deci-AI/super-gradients/pull/820)
- Introduce min_samples param to dataloader_params, which repeats the dataset in case its size is smaller than the value of this parameter. PR: [838](https://github.com/Deci-AI/super-gradients/pull/838)


Bug fixes:
- Bug Fix wandb logger on resume by [Louis-Dupont](https://github.com/Louis-Dupont) in [#766](https://github.com/Deci-AI/super-gradients/pull/766)

Bug fixes:
Fix the compatibility of ONNX conversion with torch 1.11+
Fix the docker publication
Fix Pre-Prediciton callback override bug
Fix dataloader multiprocessing (for Mac OS and Windows)
Fix strict_load not being passed in evaluate_from_recipe
Resolve many potential circular imports
Correct albumentations import log
Fix incorrect tensor type & device when entire batch has no targets
Fix documentation format in many files

Features:
New documentation for many features (EMA, AMP, QAT, Knowledge-Distillation, Dataloaders, Optimizer, Schedulers…)
Added Lion optimizer
Added registration for SG Logger (to allow custom-defined loggers to be used from YAML)
DDRNet 39 pre-trained segmentation model
Mapillary dataset
RoboFlow100 benchmark and all 100 datasets
QATTrainer for training with QAT
QAT from recipe script
PP-YoloE pre-trained models
Pose Estimation models loss functions
Added Auxiliary heads to Unet
Channel wise distillation
ONNX Simplifier on ONNX conversion
Unet and Loss function for binary segmentation
New Transforms: DetectionPadToSize, DetectionImagePermute
Allow activation factory type to take resolved type as inp
New strict_load mode: key_matching.