ultralytics 8.0.77 Ray[Tune] for hyperparameter optimization (#2014)

Co-authored-by: JF Chen <k-2feng@hotmail.com>
Co-authored-by: Ayush Chaurasia <ayush.chaurarsia@gmail.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

ultralytics/yolo/utils/callbacks/base.py

@@ -154,9 +154,11 @@ def add_integration_callbacks(instance):
     from .comet import callbacks as comet_callbacks
     from .hub import callbacks as hub_callbacks
     from .mlflow import callbacks as mf_callbacks
+    from .raytune import callbacks as tune_callbacks
     from .tensorboard import callbacks as tb_callbacks
+    from .wb import callbacks as wb_callbacks
 
-    for x in clearml_callbacks, comet_callbacks, hub_callbacks, tb_callbacks, mf_callbacks:
+    for x in clearml_callbacks, comet_callbacks, hub_callbacks, tb_callbacks, mf_callbacks, tune_callbacks, wb_callbacks:
         for k, v in x.items():
             if v not in instance.callbacks[k]:  # prevent duplicate callbacks addition
                 instance.callbacks[k].append(v)  # callback[name].append(func)

ultralytics/yolo/utils/callbacks/raytune.py

@@ -0,0 +1,17 @@
+try:
+    import ray
+    from ray import tune
+    from ray.air import session
+except (ImportError, AssertionError):
+    tune = None
+
+
+def on_fit_epoch_end(trainer):
+    if ray.tune.is_session_enabled():
+        metrics = trainer.metrics
+        metrics['epoch'] = trainer.epoch
+        session.report(metrics)
+
+
+callbacks = {
+    'on_fit_epoch_end': on_fit_epoch_end, } if tune else {}

ultralytics/yolo/utils/callbacks/wb.py

@@ -0,0 +1,48 @@
+# Ultralytics YOLO 🚀, GPL-3.0 license
+
+from ultralytics.yolo.utils.torch_utils import get_flops, get_num_params
+
+try:
+    import wandb as wb
+
+    assert hasattr(wb, '__version__')
+except (ImportError, AssertionError):
+    wb = None
+
+
+def on_pretrain_routine_start(trainer):
+    wb.init(project=trainer.args.project or 'YOLOv8', name=trainer.args.name, config=vars(
+        trainer.args)) if not wb.run else wb.run
+
+
+def on_fit_epoch_end(trainer):
+    wb.run.log(trainer.metrics, step=trainer.epoch + 1)
+    if trainer.epoch == 0:
+        model_info = {
+            'model/parameters': get_num_params(trainer.model),
+            'model/GFLOPs': round(get_flops(trainer.model), 3),
+            'model/speed(ms)': round(trainer.validator.speed['inference'], 3)}
+        wb.run.log(model_info, step=trainer.epoch + 1)
+
+
+def on_train_epoch_end(trainer):
+    wb.run.log(trainer.label_loss_items(trainer.tloss, prefix='train'), step=trainer.epoch + 1)
+    wb.run.log(trainer.lr, step=trainer.epoch + 1)
+    if trainer.epoch == 1:
+        wb.run.log({f.stem: wb.Image(str(f))
+                    for f in trainer.save_dir.glob('train_batch*.jpg')},
+                   step=trainer.epoch + 1)
+
+
+def on_train_end(trainer):
+    art = wb.Artifact(type='model', name=f'run_{wb.run.id}_model')
+    if trainer.best.exists():
+        art.add_file(trainer.best)
+        wb.run.log_artifact(art)
+
+
+callbacks = {
+    'on_pretrain_routine_start': on_pretrain_routine_start,
+    'on_train_epoch_end': on_train_epoch_end,
+    'on_fit_epoch_end': on_fit_epoch_end,
+    'on_train_end': on_train_end} if wb else {}