* Set `torch.cuda.manual_seed_all()` for DDP

* Update general.py

* Update general.py

* Add generator and worker seed

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update dataloaders.py

* Update dataloaders.py

* Update dataloaders.py
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Resolves silent outputs after model load

Otherwise, an error is thrown if the tensorboard logger is not included.

* upgrade onnxsim to v0.4.1
Signed-off-by: daquexian <daquexian566@gmail.com>

* Update export.py

* Update export.py

* Update export.py

* Update export.py

* Update export.py
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Resolves https://github.com/ultralytics/yolov5/issues/8623

* assert torch!=1.12.0 for DDP training

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ciCo-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Refactor optimizer initialization

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update train.py

* Update train.py
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Update ci-testing with `pip install wheel`

* Update ci-testing.yml

* Update dockerfiles

* Update `tqdm>=4.64.0` and `thop>=0.1.1`

* Update requirements.txt

* explicitly set weight_decay value

The default weight_decay value of AdamW is 1e-2, so we should set it to zero.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Cleanup
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Raise error on invalid model

Update requirements.txt

Protected divides in IOU function to resolve https://github.com/ultralytics/yolov5/issues/8539

* Fix LoadImages with dataset yaml lists

* Update dataloaders.py

* Update dataloaders.py

* Simplify/refactor PR

* Update dataloaders.py
Co-authored-by: Colin Wong <noreply@brains4drones.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Improved error reporting for https://github.com/ultralytics/yolov5/issues/8519

Properly expose `batch_size` from OpenVINO
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

* Add `--hard-fail` list argument to benchmarks for CI

Will cause CI to fail on a benchmark failure for given indices.

* Update ci-testing.yml

* Attempt Failure (CI should fail)

* Update benchmarks.py

* Update export.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update benchmarks.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update ci-testing.yml

* Update benchmarks.py

* Update benchmarks.py
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Avoid FP64 ops for MPS support

Resolves https://github.com/ultralytics/yolov5/pull/7878#issuecomment-1177952614

Revert "Expose OpenVINO `batch_size` similarly to TensorRT (#8437)"

This reverts commit da2ee393.

* attempt at reproducibility

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* use deterministic algs

* fix everything :)

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* revert dataloader changes

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* process_batch as np

* remove newline

* Remove dataloader init fcn

* Update val.py

* Update train.py

* revert additional changes

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update train.py

* Add --seed arg

* Update general.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update train.py

* Update train.py

* Update val.py

* Update train.py

* Update general.py

* Update general.py

* Add deterministic argument to init_seeds()
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Exclude torch==1.12.0, torchvision==0.13.0

updates:
- [github.com/pre-commit/pre-commit-hooks: v4.2.0 → v4.3.0](https://github.com/pre-commit/pre-commit-hooks/compare/v4.2.0...v4.3.0)
- [github.com/asottile/pyupgrade: v2.32.1 → v2.34.0](https://github.com/asottile/pyupgrade/compare/v2.32.1...v2.34.0)
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Require explicit request for MPS, i.e.
```bash
python detect.py --device mps
```

Reverts https://github.com/ultralytics/yolov5/pull/8210 for preferring MPS if available. 

Note that torch MPS is experiencing ongoing compatibility issues in https://github.com/pytorch/pytorch/issues/77886