Adding the MiVOLOv1 Face only estimator. Minor refactoring.

src/facebias/estimators/mivolo/mi_volo.py

@@ -217,8 +217,9 @@ class MiVOLO:
     def inference(self, model_input: torch.Tensor) -> torch.Tensor:
         with torch.no_grad():
             if self.half:
-                model_input = model_input.half()
+                output = self.model(model_input.half())
-            output = self.model(model_input)
+            else:
+                output = self.model(model_input)
         return output
 
     # def predict(self, image: np.ndarray, detected_bboxes: PersonAndFaceResult):
@@ -245,42 +246,7 @@ class MiVOLO:
     #     # write gender and age results into detected_bboxes
     #     self.fill_in_results(output, detected_bboxes, faces_inds, bodies_inds)
 
-    def fill_in_results(self, output, detected_bboxes, faces_inds, bodies_inds):
+    # def prepare_crops(self, image: np.ndarray, detected_bboxes: PersonAndFaceResult):
-        if self.meta.only_age:
-            age_output = output
-            gender_probs, gender_indx = None, None
-        else:
-            age_output = output[:, 2]
-            gender_output = output[:, :2].softmax(-1)
-            gender_probs, gender_indx = gender_output.topk(1)
-
-        assert output.shape[0] == len(faces_inds) == len(bodies_inds)
-
-        # per face
-        for index in range(output.shape[0]):
-            face_ind = faces_inds[index]
-            body_ind = bodies_inds[index]
-
-            # get_age
-            age = age_output[index].item()
-            age = age * (self.meta.max_age - self.meta.min_age) + self.meta.avg_age
-            age = round(age, 2)
-
-            detected_bboxes.set_age(face_ind, age)
-            detected_bboxes.set_age(body_ind, age)
-
-            _logger.info(f"\tage: {age}")
-
-            if gender_probs is not None:
-                gender = "male" if gender_indx[index].item() == 0 else "female"
-                gender_score = gender_probs[index].item()
-
-                _logger.info(f"\tgender: {gender} [{int(gender_score * 100)}%]")
-
-                detected_bboxes.set_gender(face_ind, gender, gender_score)
-                detected_bboxes.set_gender(body_ind, gender, gender_score)
-
-    # def prepare_crops_old(self, image: np.ndarray, detected_bboxes: PersonAndFaceResult):
 
     #     crops: PersonAndFaceCrops = detected_bboxes.collect_crops(image)
     #     (bodies_inds, bodies_crops), (faces_inds, faces_crops) = crops.get_faces_with_bodies(

src/facebias/estimators/mivolov1.py

@@ -0,0 +1,95 @@
+# -*- coding: utf-8 -*-
+
+from collections import OrderedDict
+from pathlib import Path
+from typing import Any, Optional
+
+import numpy as np
+import torch
+import torch.nn as nn
+import torchvision
+from facebias import load_dataset
+from facebias.estimators import BaseEstimator, Capability
+from facebias.estimators.mivolo.mi_volo import MiVOLO
+from timm.data import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
+from torchvision import transforms
+
+
+class MiVOLOv1(BaseEstimator):
+    """Generic wrapper for a local MiVOLO v1-style face only model.
+
+    Parameters
+    ----------
+    checkpoint_path: Path
+        Path to the saved checkpoint. Note that we expect the checkpoint to be
+        in PyTorch format.
+
+    device: Optional[torch.device]
+        Device to load the checkpoints into. By default is "cpu".
+    """
+    def __init__(
+            self,
+            checkpoint_path: Path,
+            device: Optional[torch.device]=torch.device("cpu")
+    ):
+        self.T = transforms.Compose([
+            transforms.ToPILImage(),
+            transforms.Resize((224, 224)),
+            transforms.ToTensor(),
+            transforms.Normalize(mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD),
+        ])
+
+        self.device = device
+        self.model = MiVOLO(checkpoint_path, half=False, device=device)
+
+    def predict(self, ims: dict[Path, np.ndarray]) -> dict[Path, dict[Capability, Any]]:
+        preds = OrderedDict()
+        min_age = self.model.meta.min_age
+        max_age = self.model.meta.max_age
+        avg_age = self.model.meta.avg_age
+
+        for p, im in ims.items():
+            im = self.T(im).view(1, 3, 224, 224).to(self.device)
+            with torch.no_grad():
+                y = self.model.inference(im)
+
+            y = y.cpu().detach()
+
+            # Age calc.
+            y_age = y[:, 2]
+            age = y_age.item() * (max_age - min_age) + avg_age
+            age = round(age, 2)
+
+            # Gender est.
+            y_gender = y[:, :2].softmax(-1)
+            gender_prob, gender_idx = y_gender.topk(1)
+            gender_label = 'm' if gender_idx.item() == 0 else 'f'
+
+            preds[p] = {
+                Capability.AGE: age,
+                Capability.AGEGROUP: _to_age_label(age),
+                Capability.SEX: gender_label,
+            }
+
+        return preds
+
+    def capabilities() -> list[Capability]:
+        return [Capability.AGE, Capability.AGEGROUP, Capability.SEX]
+
+
+def _to_age_label(age):
+    base_age = int(age // 10 * 10)
+    return "{}-{}".format(base_age, base_age + 9)
+
+
+if __name__ == '__main__':
+    from facebias import load_dataset
+
+    DATASET_PATH = Path("data/frll_neutral_front")
+    dataset, _ = load_dataset(DATASET_PATH, imname_proc_fn=lambda x: x.split('_')[0])
+
+    print(MiVOLOv1.capabilities())
+    MODEL_PATH = Path("models/volo-v1_model_imdb_age_gender_4.22.pth.tar")
+    model = MiVOLOv1(MODEL_PATH)
+
+    out = model.predict(dataset)