Netspresso-trainer

New Features:

- Update Benchmarks & Checkpoints (docs) and weights files to fully usable by `illian01` in [PR 446](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/446), [PR 447](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/447), [PR 456](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/456), [PR 461](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/461)
- Add TFLite runtime code example by `illian01` in [PR 449](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/449)

Bug Fixes:

- Fix best_epoch init error in TrainingSummary in case of training resume by `illian01` in [PR 448](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/448)
- Fix segmentation metric logic bug by `illian01` in [PR 455](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/455), [PR 460](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/460)

Breaking Changes:

No changes to highlight.

Other Changes:

- Refactoring: remove thop, replace MACs with FLOPs by `illian01` in [PR 444](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/444)
- Add copyright for entire project by `illian01` in [PR 451](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/451)
- Modify ImageSaver to receive various resolution at once `illian01` in [PR 458](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/458)

New Features:

- Add dataset validation step and refactoring data modules by `illian01` in [PR 417](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/417), [PR 419](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/419)
- Add various dataset examples including automatic open dataset format converter by `illian01` in [PR 430](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/430)
- Allow using text file path for the `id_mapping` field by `illian01` in [PR 432](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/432), [PR 435](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/435)

Bug Fixes:

- Fix test directory check line by `illian01` in [PR 428](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/428)
- Fix Dockerfile installation commandline `cbpark-nota` in [PR 434](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/434)

Breaking Changes:

No changes to highlight.

Other Changes:

- Save training summary at every end of epochs by `illian01` in [PR 420](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/420)
- Refacotring: rename postprocessors/register.py to registry.py by `aychun` in [PR 424](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/424)
- Add example configuration set by `illian01` in [PR 438](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/438)
- Documentation: fix simple use config file path by `cbpark-nota` in [PR 437](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/437)

New Features:

- Add activation and dropout layer in FC by `illian01` in [PR 325](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/325), [PR 327](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/327)
- Add function to Resize: Match longer side with input size and keep ratio by `illian01` in [PR 329](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/329)
- Add transforms: MosaicDetection by `illian01` in [PR 331](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/331), [PR 337](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/337), [PR 397](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/397)
- Add transform: HSVJitter by `illian01` in [PR 336](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/336), [PR 413](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/413)
- Add transforms: RandomResize by `illian01` in [PR 341](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/341), [PR 344](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/344), [PR 398](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/398)
- Add model EMA (Exponential Moving Average) by `illian01` in [PR 348](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/348)
- Add entry point for evaluation and inference by `illian01` in [PR 374](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/374), [PR 379](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/379), [PR 381](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/381), [PR 383](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/383)
- Add classification visulizer by `illian01` in [PR 384](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/384)
- Add dataset caching feature by `illian01` in [PR 391](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/391)
- Add mixed precision training by `illian01` in [PR 392](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/392)
- Add YOLOX l1 loss activation option by `illian01` in [PR 396](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/396)
- Add NetsPresso Trainer YOLOX pretrained weights by `illian01` in [PR 406](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/406)

Bug Fixes:

- Fix output_root_dir from fixed string to config value by `illian01` in [PR 323](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/323)
- Gather predicted results before compute metric and fix additional distributed evaluation inaccurate error by `illian01` in [PR 346](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/346), [PR 356](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/356)
- Fix detection score return by `illian01` in [PR 373](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/373)
- Fix memory leak from onnx export by `illian01` in [PR 386](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/386), [PR 394](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/394)
- Refactoring metric modules and fix inaccurate metric bug by `illian01` in [PR 402](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/402)

Breaking Changes:

- Simplify augmentation configuration hierarchy by `illian01` in [PR 322](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/322)
- Add pose estimation task and RTMPose model by `illian01` in [PR 357](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/357), [PR 366](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/366)
- Remove pythonic config and move training initialization functions to `trainer_main.py` by `illian01` in [PR 371](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/371)
- Unify gradio demo in one page by `deepkyu` in [PR 408](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/408)

Other Changes:

- Refactoring: Move custom transforms to each python module by `illian01` in [PR 332](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/332)
- Update Pad transform to receive target size of image by `illian01` in [PR 334](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/334)
- Rafactoring: Fix to make transform object in init by `illian01` in [PR 339](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/339)
- ~~Add before_epoch step which does update modules like dataloader before epoch training by `illian01` in [PR 340](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/340)~~
- Revert [PR 340](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/340) and add multiprocessing.Value to handle MosaicDetection and RandomResize by `illian01` in [PR 345](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/345)
- Enable adjust max epoch of scheduler by `illian01` in [PR 350](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/350)
- Remove github action about hugging face space demo by `illian01` in [PR 351](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/351)
- Update docs by `illian01` in [PR 355](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/355), [PR 410](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/410)
- Backbone task compatibility checking refactoring by `illian01` in [PR 361](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/361), [PR 364](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/364)
- Fix postprocessor return type as numpy.ndarray by `illian01` in [PR 365](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/365)
- Update default asignees of issue template by `illian01` in [PR 375](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/375)
- Refactoring: Remove CSV logger, change logger module input format by `illian01` in [PR 377](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/377)
- Change ClassficationDataSampler logic by `illian01` in [PR 382](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/382)
- Add YOLOX weights initialization step by `illian01` in [PR 393](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/393)
- Minor update: detection postprocessor, dataset, and padding strategy by `illian01` in [PR 395](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/395)
- Specify input size for onnx export and remove augmentation.img_size by `illian01` in [PR 399](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/399)
- Update issue and pr template by `illian01` in [PR 401](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/401)
- Add documentation auto deploy action by `illian01` in [PR 405](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/405)

New Features:

No changes to highlight.

Bug Fixes:

- Remove union of int and list by `illian01` in [PR 317](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/317)

Breaking Changes:

No changes to highlight.

Other Changes:

No changes to highlight.

New Features:

- Enable customizing inference transform by `illian01` in [PR 304](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/304)
- Add transform function: CenterCrop by `illian01` in [PR 308](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/308)

Bug Fixes:

- Fix automatic PIDNet weights download bug by `illian01` in [PR 306](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/306)
- Resize default value to list by `illian01` in [PR 315](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/315)

Breaking Changes:

No changes to highlight.

Other Changes:

- Update model caching directory and checkpoint configuration by `deepkyu` in [PR 299](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/299), [PR 312](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/312)
- Minor docs update by `illian01` in [PR 300](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/300)
- Update software development stage by `illian01` in [PR 301](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/301)
- Fix size param of Resize to receive int or list by `illian01` in [PR 310](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/310)
- Modify PIDNet conv bias, add head_list property on models by `illian01` in [PR 311](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/311)

New Features:

- Construct head by config file by `illian01` in [PR 237](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/237)
- Construct neck by config file by `illian01` in [PR 249](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/249)
- Add model: RetinaNet by `illian01` in [PR 257](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/257)
- Select `gpus` with `environment` configuration by `deepkyu` in [PR 269](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/269)
- Return logging directory path and fix training interfaces by `deepkyu` in [PR 271](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/271)
- Add transform: AutoAugment by `illian01` in [PR 281](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/281)

Bug Fixes:

- Fix attribute error on fc by `illian01` in [PR 252](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/252)
- Restore file export for stream log by `deepkyu` in [PR 255](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/255)
- Fix CSV logging, configuration error, and misused loggings by `deepkyu` in [PR 259](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/259)
- Fix minor bug in train.py by `illian01` in [PR 277](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/277)
- Fix local classification dataset loader error by `illian01` in [PR 279](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/279)
- Fix safetensors file overwriting bug by `illian01` in [PR 289](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/289)
- Fix error on full model load by `illian01` in [PR 295](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/295)

Breaking Changes:

- Provide pytorch state dict with `.safetensors` and training summary with `.json` for a better utilization by `deepkyu` in [PR 262](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/262)

Other Changes:

- Refactoring for detection models by `illian01` in [PR 260](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/260)
- Equalize logging format with `PyNetsPresso` by `deepkyu` in [PR 263](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/263)
- Refactoring for clean docs by `illian01` in [PR 265](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/265), [PR 266](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/266), [PR 272](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/272), [PR 273](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/273), [PR 274](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/274), [PR 284](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/284)
- Update docs up-to-date by `illian01` in [PR 278](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/278)
- Refactoring model building code and move TASK_MODEL_DICT by `illian01` in [PR 282](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/282)
- Add eps param on RMSprop by `illian01` in [PR 285](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/285)
- Fix weights loading logic by `illian01` in [PR 287](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/287), [PR 290](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/290)
- Change pretrained checkpoint name convention and update weight path and url by `illian01` in [PR 291](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/291)
- Move seed field to environment config by `illian01` in [PR 292](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/292)
- Move ResNet and Fc implementation code to core directory by `illian01` in [PR 294](https://github.com/Nota-NetsPresso/netspresso-trainer/pull/294)