Add MNv4 medium recipe

CHANGELOG.md

@@ -4,6 +4,11 @@ All notable changes to this project will be documented in this file.
 
 ## \[Unreleased\]
 
+### New features
+
+- Add MobileNetV4 medium recipe
+  (https://github.com/open-edge-platform/training_extensions/pull/4341)
+
 ### Enhancements
 
 - Bump OV and NNCF to 2025.1

README.md

@@ -15,7 +15,7 @@
 
 [![python](https://img.shields.io/badge/python-3.10%2B-green)]()
 [![pytorch](https://img.shields.io/badge/pytorch-2.5%2B-orange)]()
-[![openvino](https://img.shields.io/badge/openvino-2025.0-purple)]()
+[![openvino](https://img.shields.io/badge/openvino-2025.1-purple)]()
 
 <!-- markdownlint-enable  MD042 -->
 

src/otx/engine/utils/auto_configurator.py

@@ -48,8 +48,8 @@
 }
 
 TASK_PER_DATA_FORMAT = {
-    "imagenet_with_subset_dirs": [OTXTaskType.MULTI_CLASS_CLS, OTXTaskType.H_LABEL_CLS],
-    "datumaro": [OTXTaskType.MULTI_LABEL_CLS],
+    "imagenet_with_subset_dirs": [OTXTaskType.MULTI_CLASS_CLS],
+    "datumaro": [OTXTaskType.MULTI_LABEL_CLS, OTXTaskType.H_LABEL_CLS],
     "coco_person_keypoints": [OTXTaskType.KEYPOINT_DETECTION],
     "coco_instances": [
         OTXTaskType.DETECTION,

src/otx/recipe/classification/h_label_cls/mobilenet_v4_medium.yaml

@@ -0,0 +1,42 @@
+model:
+  class_path: otx.algo.classification.hlabel_models.timm_model.TimmModelHLabelCls
+  init_args:
+    model_name: mobilenetv4_medium_conv
+
+    optimizer:
+      class_path: torch.optim.SGD
+      init_args:
+        lr: 0.0058
+        momentum: 0.9
+        weight_decay: 0.0001
+
+    scheduler:
+      class_path: otx.core.schedulers.LinearWarmupSchedulerCallable
+      init_args:
+        num_warmup_steps: 10
+        main_scheduler_callable:
+          class_path: lightning.pytorch.cli.ReduceLROnPlateau
+          init_args:
+            mode: max
+            factor: 0.5
+            patience: 3
+            monitor: val/accuracy
+
+engine:
+  task: H_LABEL_CLS
+  device: auto
+
+callback_monitor: val/accuracy
+
+data: ../../_base_/data/classification.yaml
+overrides:
+  max_epochs: 90
+
+  callbacks:
+    - class_path: otx.algo.callbacks.adaptive_early_stopping.EarlyStoppingWithWarmup
+      init_args:
+        patience: 5
+
+  data:
+    task: H_LABEL_CLS
+    data_format: datumaro

src/otx/recipe/classification/multi_class_cls/mobilenet_v4.yaml → src/otx/recipe/classification/multi_class_cls/mobilenet_v4_medium.yaml

@@ -1,14 +1,17 @@
 model:
   class_path: otx.algo.classification.multiclass_models.timm_model.TimmModelMulticlassCls
   init_args:
-    model_name: mobilenetv4_conv_small
+    model_name: mobilenetv4_medium_conv
     label_info: 1000
+    freeze_backbone: False
+
     optimizer:
       class_path: torch.optim.SGD
       init_args:
         lr: 0.0058
         momentum: 0.9
         weight_decay: 0.0001
+
     scheduler:
       class_path: otx.core.schedulers.LinearWarmupSchedulerCallable
       init_args:
@@ -20,13 +23,17 @@ model:
             factor: 0.5
             patience: 3
             monitor: val/accuracy
+
 engine:
   task: MULTI_CLASS_CLS
   device: auto
+
 callback_monitor: val/accuracy
-data: ../../base/data/classification.yaml
+
+data: ../../_base_/data/classification.yaml
 overrides:
   max_epochs: 90
+
   callbacks:
     - class_path: otx.algo.callbacks.adaptive_early_stopping.EarlyStoppingWithWarmup
       init_args:

src/otx/recipe/classification/multi_label_cls/mobilenet_v4_medium.yaml

@@ -0,0 +1,41 @@
+model:
+  class_path: otx.algo.classification.multilabel_models.timm_model.TimmModelMultilabelCls
+  init_args:
+    model_name: mobilenetv3_large
+    label_info: 1000
+
+    optimizer:
+      class_path: torch.optim.SGD
+      init_args:
+        lr: 0.0058
+        momentum: 0.9
+        weight_decay: 0.0005
+
+    scheduler:
+      class_path: otx.core.schedulers.LinearWarmupSchedulerCallable
+      init_args:
+        num_warmup_steps: 10
+        main_scheduler_callable:
+          class_path: lightning.pytorch.cli.ReduceLROnPlateau
+          init_args:
+            mode: max
+            factor: 0.5
+            patience: 3
+            monitor: val/accuracy
+
+engine:
+  task: MULTI_LABEL_CLS
+  device: auto
+
+callback_monitor: val/accuracy
+
+data: ../../_base_/data/classification.yaml
+overrides:
+  data:
+    task: MULTI_LABEL_CLS
+    data_format: datumaro
+
+  callbacks:
+    - class_path: otx.algo.callbacks.adaptive_early_stopping.EarlyStoppingWithWarmup
+      init_args:
+        patience: 5