refactor(retrieval): polish retrieval.py

ppcls/engine/evaluation/retrieval.py

@@ -16,108 +16,90 @@ from __future__ import division
 from __future__ import print_function
 
 import platform
+from collections import defaultdict
 
 import numpy as np
 import paddle
-from ppcls.engine.train.utils import type_name
-from ppcls.utils import logger
-from ppcls.utils import all_gather
+
+from ppcls.utils import all_gather, logger
 
 
 def retrieval_eval(engine, epoch_id=0):
     engine.model.eval()
     # step1. prepare query and gallery features
     if engine.gallery_query_dataloader is not None:
-        gallery_feas, gallery_label_id, gallery_camera_id = compute_feature(
+        gallery_feat, gallery_label, gallery_camera = compute_feature(
             engine, "gallery_query")
-        query_feas, query_label_id, query_camera_id = gallery_feas, gallery_label_id, gallery_camera_id
+        query_feat, query_label, query_camera = gallery_feat, gallery_label, gallery_camera
     else:
-        gallery_feas, gallery_label_id, gallery_camera_id = compute_feature(
+        gallery_feat, gallery_label, gallery_camera = compute_feature(
             engine, "gallery")
-        query_feas, query_label_id, query_camera_id = compute_feature(engine,
-                                                                      "query")
-
+        query_feat, query_label, query_camera = compute_feature(engine,
+                                                                "query")
     # step2. split features into feature blocks for saving memory
+    num_query = len(query_feat)
     block_size = engine.config["Global"].get("sim_block_size", 64)
-    sections = [block_size] * (len(query_feas) // block_size)
-    if len(query_feas) % block_size > 0:
-        sections.append(len(query_feas) % block_size)
-
-    query_feas_blocks = paddle.split(query_feas, sections)
-    query_camera_id_blocks = (paddle.split(query_camera_id, sections)
-                              if query_camera_id is not None else None)
-    query_label_id_blocks = paddle.split(query_label_id, sections)
+    sections = [block_size] * (num_query // block_size)
+    if num_query % block_size > 0:
+        sections.append(num_query % block_size)
+
+    query_feat_blocks = paddle.split(query_feat, sections)
+    query_label_blocks = paddle.split(query_label, sections)
+    query_camera_blocks = paddle.split(
+        query_camera, sections) if query_camera is not None else None
     metric_key = None
 
     # step3. compute metric
     if engine.eval_loss_func is None:
-        metric_dict = {metric_key: 0.}
+        metric_dict = {metric_key: 0.0}
     else:
         use_reranking = engine.config["Global"].get("re_ranking", False)
         logger.info(f"re_ranking={use_reranking}")
-        metric_dict = {}
         if use_reranking:
-            for _, metric_func in enumerate(
-                    engine.eval_metric_func.metric_func_list):
-                if hasattr(metric_func,
-                           "descending") and metric_func.descending is True:
-                    metric_func.descending = False
-                    logger.warning(
-                        f"re_ranking=True, set {type_name(metric_func)}.descending set to False"
-                    )
             # compute distance matrix
             distmat = compute_re_ranking_dist(
-                query_feas, gallery_feas, engine.config["Global"].get(
+                query_feat, gallery_feat, engine.config["Global"].get(
                     "feature_normalize", True), 20, 6, 0.3)
-
             # exclude illegal distance
-            camera_id_mask = query_camera_id != gallery_camera_id.t()
-            image_id_mask = query_label_id != gallery_label_id.t()
-            keep_mask = paddle.logical_or(image_id_mask, camera_id_mask)
-            distmat = distmat * keep_mask.astype(query_feas.dtype)
-            inf_mat = (
-                paddle.logical_not(keep_mask).astype(query_feas.dtype)) * (
-                    distmat.max() + 1)
-            distmat = distmat + inf_mat
-
-            metric_block = engine.eval_metric_func(distmat, query_label_id,
-                                                   gallery_label_id, keep_mask)
-            for key in metric_block:
-                metric_dict[key] = metric_block[key]
+            if query_camera is not None:
+                camera_mask = query_camera != gallery_camera.t()
+                label_mask = query_label != gallery_label.t()
+                keep_mask = label_mask | camera_mask
+                distmat = keep_mask.astype(query_feat.dtype) * distmat + (
+                    ~keep_mask).astype(query_feat.dtype) * (distmat.max() + 1)
+            else:
+                keep_mask = None
+            # compute metric with all samples
+            metric_dict = engine.eval_metric_func(-distmat, query_label,
+                                                  gallery_label, keep_mask)
         else:
-            for block_idx, block_fea in enumerate(query_feas_blocks):
+            metric_dict = defaultdict(float)
+            for block_idx, block_feat in enumerate(query_feat_blocks):
+                # compute distance matrix
                 distmat = paddle.matmul(
-                    block_fea, gallery_feas, transpose_y=True)
-                if query_camera_id is not None:
-                    query_camera_id_block = query_camera_id_blocks[block_idx]
-                    camera_id_mask = query_camera_id_block != gallery_camera_id.t(
-                    )
-
-                    query_label_id_block = query_label_id_blocks[block_idx]
-                    image_id_mask = query_label_id_block != gallery_label_id.t(
-                    )
-
-                    keep_mask = paddle.logical_or(image_id_mask,
-                                                  camera_id_mask)
-                    distmat = distmat * keep_mask.astype("float32")
+                    block_feat, gallery_feat, transpose_y=True)
+                # exclude illegal distance
+                if query_camera is not None:
+                    camera_mask = query_camera_blocks[
+                        block_idx] != gallery_camera.t()
+                    label_mask = query_label_blocks[
+                        block_idx] != gallery_label.t()
+                    keep_mask = label_mask | camera_mask
+                    distmat = keep_mask.astype(query_feat.dtype) * distmat
                 else:
                     keep_mask = None
-
+                # compute metric by block
                 metric_block = engine.eval_metric_func(
-                    distmat, query_label_id_blocks[block_idx],
-                    gallery_label_id, keep_mask)
-
+                    distmat, query_label_blocks[block_idx], gallery_label,
+                    keep_mask)
+                # accumulate metric
                 for key in metric_block:
-                    if key not in metric_dict:
-                        metric_dict[key] = metric_block[key] * block_fea.shape[
-                            0] / len(query_feas)
-                    else:
-                        metric_dict[key] += metric_block[
-                            key] * block_fea.shape[0] / len(query_feas)
+                    metric_dict[key] += metric_block[key] * block_feat.shape[
+                        0] / num_query
 
     metric_info_list = []
-    for key in metric_dict:
-        metric_info_list.append(f"{key}: {metric_dict[key]:.5f}")
+    for key, value in metric_dict.items():
+        metric_info_list.append(f"{key}: {value:.5f}")
         if metric_key is None:
             metric_key = key
     metric_msg = ", ".join(metric_info_list)
@@ -127,9 +109,6 @@ def retrieval_eval(engine, epoch_id=0):
 
 
 def compute_feature(engine, name="gallery"):
-    has_camera_id = False
-    all_camera_id = None
-
     if name == "gallery":
         dataloader = engine.gallery_dataloader
     elif name == "query":
@@ -137,13 +116,16 @@ def compute_feature(engine, name="gallery"):
     elif name == "gallery_query":
         dataloader = engine.gallery_query_dataloader
     else:
-        raise RuntimeError("Only support gallery or query dataset")
+        raise ValueError(
+            f"Only support gallery or query or gallery_query dataset, but got {name}"
+        )
 
-    batch_feas_list = []
-    label_id_list = []
-    camera_id_list = []
+    all_feat = []
+    all_label = []
+    all_camera = []
     max_iter = len(dataloader) - 1 if platform.system() == "Windows" else len(
         dataloader)
+    has_camera = False
     for idx, batch in enumerate(dataloader):  # load is very time-consuming
         if idx >= max_iter:
             break
@@ -154,8 +136,8 @@ def compute_feature(engine, name="gallery"):
 
         batch = [paddle.to_tensor(x) for x in batch]
         batch[1] = batch[1].reshape([-1, 1]).astype("int64")
-        if len(batch) == 3:
-            has_camera_id = True
+        if len(batch) >= 3:
+            has_camera = True
             batch[2] = batch[2].reshape([-1, 1]).astype("int64")
         if engine.amp and engine.amp_eval:
             with paddle.amp.auto_cast(
@@ -163,62 +145,61 @@ def compute_feature(engine, name="gallery"):
                         "flatten_contiguous_range", "greater_than"
                     },
                     level=engine.amp_level):
-                out = engine.model(batch[0], batch[1])
+                out = engine.model(batch[0])
         else:
-            out = engine.model(batch[0], batch[1])
+            out = engine.model(batch[0])
         if "Student" in out:
             out = out["Student"]
 
         # get features
         if engine.config["Global"].get("retrieval_feature_from",
                                        "features") == "features":
-            # use neck's output as features
-            batch_feas = out["features"]
+            # use output from neck as feature
+            batch_feat = out["features"]
         else:
-            # use backbone's output as features
-            batch_feas = out["backbone"]
+            # use output from backbone as feature
+            batch_feat = out["backbone"]
 
-        # do norm(optinal)
+        # do norm(optional)
         if engine.config["Global"].get("feature_normalize", True):
-            batch_feas_norm = paddle.sqrt(
-                paddle.sum(paddle.square(batch_feas), axis=1, keepdim=True))
-            batch_feas = paddle.divide(batch_feas, batch_feas_norm)
+            batch_feat = paddle.nn.functional.normalize(batch_feat, p=2)
 
-        # do binarize(optinal)
+        # do binarize(optional)
         if engine.config["Global"].get("feature_binarize") == "round":
-            batch_feas = paddle.round(batch_feas).astype("float32") * 2.0 - 1.0
+            batch_feat = paddle.round(batch_feat).astype("float32") * 2.0 - 1.0
         elif engine.config["Global"].get("feature_binarize") == "sign":
-            batch_feas = paddle.sign(batch_feas).astype("float32")
+            batch_feat = paddle.sign(batch_feat).astype("float32")
 
         if paddle.distributed.get_world_size() > 1:
-            batch_feas_list.append(all_gather(batch_feas))
-            label_id_list.append(all_gather(batch[1]))
-            if has_camera_id:
-                camera_id_list.append(all_gather(batch[2]))
+            all_feat.append(all_gather(batch_feat))
+            all_label.append(all_gather(batch[1]))
+            if has_camera:
+                all_camera.append(all_gather(batch[2]))
         else:
-            batch_feas_list.append(batch_feas)
-            label_id_list.append(batch[1])
-            if has_camera_id:
-                camera_id_list.append(batch[2])
+            all_feat.append(batch_feat)
+            all_label.append(batch[1])
+            if has_camera:
+                all_camera.append(batch[2])
 
     if engine.use_dali:
         dataloader.reset()
 
-    all_feas = paddle.concat(batch_feas_list)
-    all_label_id = paddle.concat(label_id_list)
-    if has_camera_id:
-        all_camera_id = paddle.concat(camera_id_list)
-
+    all_feat = paddle.concat(all_feat)
+    all_label = paddle.concat(all_label)
+    if has_camera:
+        all_camera = paddle.concat(all_camera)
+    else:
+        all_camera = None
     # discard redundant padding sample(s) at the end
-    total_samples = len(
-        dataloader.dataset) if not engine.use_dali else dataloader.size
-    all_feas = all_feas[:total_samples]
-    all_label_id = all_label_id[:total_samples]
-    if has_camera_id:
-        all_camera_id = all_camera_id[:total_samples]
+    total_samples = dataloader.size if engine.use_dali else len(
+        dataloader.dataset)
+    all_feat = all_feat[:total_samples]
+    all_label = all_label[:total_samples]
+    if has_camera:
+        all_camera = all_camera[:total_samples]
 
-    logger.info(f"Build {name} done, all feat shape: {all_feas.shape}")
-    return all_feas, all_label_id, all_camera_id
+    logger.info(f"Build {name} done, all feat shape: {all_feat.shape}")
+    return all_feat, all_label, all_camera
 
 
 def k_reciprocal_neighbor(rank: np.ndarray, p: int, k: int) -> np.ndarray:
@@ -239,8 +220,8 @@ def k_reciprocal_neighbor(rank: np.ndarray, p: int, k: int) -> np.ndarray:
     return forward_k_neigh_index[candidate]
 
 
-def compute_re_ranking_dist(query_feas: paddle.Tensor,
-                            gallery_feas: paddle.Tensor,
+def compute_re_ranking_dist(query_feat: paddle.Tensor,
+                            gallery_feat: paddle.Tensor,
                             feature_normed: bool=True,
                             k1: int=20,
                             k2: int=6,
@@ -251,8 +232,8 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
     Code refernence: https://github.com/michuanhaohao/reid-strong-baseline/blob/master/utils/re_ranking.py
 
     Args:
-        query_feas (paddle.Tensor): Query features with shape of [num_query, feature_dim].
-        gallery_feas (paddle.Tensor):  Gallery features with shape of [num_gallery, feature_dim].
+        query_feat (paddle.Tensor): Query features with shape of [num_query, feature_dim].
+        gallery_feat (paddle.Tensor):  Gallery features with shape of [num_gallery, feature_dim].
         feature_normed (bool, optional):  Whether input features are normalized.
         k1 (int, optional): Parameter for K-reciprocal nearest neighbors. Defaults to 20.
         k2 (int, optional): Parameter for K-nearest neighbors. Defaults to 6.
@@ -261,10 +242,10 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
     Returns:
         paddle.Tensor: (1 - lamb) x Dj + lamb x D, with shape of [num_query, num_gallery].
     """
-    num_query = query_feas.shape[0]
-    num_gallery = gallery_feas.shape[0]
+    num_query = query_feat.shape[0]
+    num_gallery = gallery_feat.shape[0]
     num_all = num_query + num_gallery
-    feat = paddle.concat([query_feas, gallery_feas], 0)
+    feat = paddle.concat([query_feat, gallery_feat], 0)
     logger.info("Using GPU to compute original distance matrix")
 
     # use L2 distance
@@ -273,8 +254,7 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
     else:
         original_dist = paddle.pow(feat, 2).sum(axis=1, keepdim=True).expand([num_all, num_all]) + \
             paddle.pow(feat, 2).sum(axis=1, keepdim=True).expand([num_all, num_all]).t()
-        original_dist = original_dist.addmm(
-            x=feat, y=feat.t(), alpha=-2.0, beta=1.0)
+        original_dist = original_dist.addmm(feat, feat.t(), -2.0, 1.0)
     original_dist = original_dist.numpy()
     del feat
 
@@ -298,7 +278,6 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
                 p_k_reciprocal_exp_ind = np.append(p_k_reciprocal_exp_ind,
                                                    q_k_reciprocal_ind)
         p_k_reciprocal_exp_ind = np.unique(p_k_reciprocal_exp_ind)
-
         # reweight distance using gaussian kernel
         weight = np.exp(-original_dist[p, p_k_reciprocal_exp_ind])
         V[p, p_k_reciprocal_exp_ind] = weight / np.sum(weight)
@@ -318,6 +297,7 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
     for gj in range(num_all):
         invIndex.append(np.nonzero(V[:, gj])[0])
 
+    # compute jaccard distance
     jaccard_dist = np.zeros_like(original_dist, dtype=np.float16)
     for p in range(num_query):
         sum_min = np.zeros(shape=[1, num_all], dtype=np.float16)
@@ -328,7 +308,8 @@ def compute_re_ranking_dist(query_feas: paddle.Tensor,
             sum_min[0, gi] += np.minimum(V[p, gj], V[gi, gj])
         jaccard_dist[p] = 1 - sum_min / (2 - sum_min)
 
-    final_dist = jaccard_dist * (1 - lamb) + original_dist * lamb
+    # fuse jaccard distance with original distance
+    final_dist = (1 - lamb) * jaccard_dist + lamb * original_dist
     del original_dist
     del V
     del jaccard_dist

ppcls/metric/metrics.py

@@ -287,10 +287,10 @@ class Recallk(nn.Layer):
                 keep_mask):
         metric_dict = dict()
 
-        #get cmc
+        # get cmc
         choosen_indices = paddle.argsort(
             similarities_matrix, axis=1, descending=self.descending)
-        gallery_labels_transpose = paddle.transpose(gallery_img_id, [1, 0])
+        gallery_labels_transpose = gallery_img_id.t()
         gallery_labels_transpose = paddle.broadcast_to(
             gallery_labels_transpose,
             shape=[
@@ -301,18 +301,14 @@ class Recallk(nn.Layer):
         equal_flag = paddle.equal(choosen_label, query_img_id)
         if keep_mask is not None:
             keep_mask = paddle.index_sample(
-                keep_mask.astype('float32'), choosen_indices)
-            equal_flag = paddle.logical_and(equal_flag,
-                                            keep_mask.astype('bool'))
-        equal_flag = paddle.cast(equal_flag, 'float32')
+                keep_mask.astype("float32"), choosen_indices)
+            equal_flag = equal_flag & keep_mask.astype("bool")
+        equal_flag = paddle.cast(equal_flag, "float32")
         real_query_num = paddle.sum(equal_flag, axis=1)
-        real_query_num = paddle.sum(
-            paddle.greater_than(real_query_num, paddle.to_tensor(0.)).astype(
-                "float32"))
+        real_query_num = paddle.sum((real_query_num > 0.0).astype("float32"))
 
         acc_sum = paddle.cumsum(equal_flag, axis=1)
-        mask = paddle.greater_than(acc_sum,
-                                   paddle.to_tensor(0.)).astype("float32")
+        mask = (acc_sum > 0.0).astype("float32")
         all_cmc = (paddle.sum(mask, axis=0) / real_query_num).numpy()
 
         for k in self.topk: