diff --git a/eval/mot/motmetrics/distances.py b/eval/mot/motmetrics/distances.py index 724dca9..5e73de5 100644 --- a/eval/mot/motmetrics/distances.py +++ b/eval/mot/motmetrics/distances.py @@ -76,7 +76,7 @@ def boxiou(a, b): a_vol = np.prod(a_size, axis=-1) b_vol = np.prod(b_size, axis=-1) u_vol = a_vol + b_vol - i_vol - return np.where(i_vol == 0, np.zeros_like(i_vol, dtype=np.float), + return np.where(i_vol == 0, np.zeros_like(i_vol, dtype=np.float32), math_util.quiet_divide(i_vol, u_vol)) diff --git a/eval/posetrack21/posetrack21/trackeval/datasets/posetrack.py b/eval/posetrack21/posetrack21/trackeval/datasets/posetrack.py index a87b817..9657982 100644 --- a/eval/posetrack21/posetrack21/trackeval/datasets/posetrack.py +++ b/eval/posetrack21/posetrack21/trackeval/datasets/posetrack.py @@ -332,7 +332,7 @@ def get_preprocessed_seq_data(self, raw_data, cls): for t in range(raw_data['num_timesteps']): if len(data['gt_ids'][t]) > 0: data['original_gt_ids'][t] = data['gt_ids'][t].copy() - data['gt_ids'][t] = gt_id_map[data['gt_ids'][t]].astype(np.int) + data['gt_ids'][t] = gt_id_map[data['gt_ids'][t]].astype(np.int32) gt_dets = data['gt_dets'][t] num_gt_joints += count_valid_joints(gt_dets) @@ -344,7 +344,7 @@ def get_preprocessed_seq_data(self, raw_data, cls): for t in range(raw_data['num_timesteps']): if len(data['tracker_ids'][t]) > 0: data['original_tracker_ids'][t] = data['tracker_ids'][t].copy() - data['tracker_ids'][t] = tracker_id_map[data['tracker_ids'][t]].astype(np.int) + data['tracker_ids'][t] = tracker_id_map[data['tracker_ids'][t]].astype(np.int32) tracker_dets = data['tracker_dets'][t] num_tracker_joints += count_valid_joints(tracker_dets) diff --git a/eval/posetrack21/posetrack21/trackeval/datasets/posetrack_mot.py b/eval/posetrack21/posetrack21/trackeval/datasets/posetrack_mot.py index 30e4401..1b57427 100644 --- a/eval/posetrack21/posetrack21/trackeval/datasets/posetrack_mot.py +++ b/eval/posetrack21/posetrack21/trackeval/datasets/posetrack_mot.py @@ -202,7 +202,7 @@ def _load_raw_file(self, tracker, seq, is_gt): time_key = str(t+1) if time_key in read_data.keys(): try: - time_data = np.asarray(read_data[time_key], dtype=np.float) + time_data = np.asarray(read_data[time_key], dtype=np.float32) except ValueError: if is_gt: raise TrackEvalException( @@ -330,7 +330,7 @@ def get_preprocessed_seq_data(self, raw_data, cls): # Match tracker and gt dets (with hungarian algorithm) and remove tracker dets which match with gt dets # which are labeled as belonging to a distractor class. - to_remove_tracker = np.array([], np.int) + to_remove_tracker = np.array([], np.int32) if gt_ids.shape[0] > 0 and tracker_ids.shape[0] > 0: # Check all classes are valid: @@ -420,7 +420,7 @@ def get_preprocessed_seq_data(self, raw_data, cls): det_idx = unmatched_det_idxs[remove_idx] dets_to_remove.append(det_idx) - to_remove_tracker = np.array(dets_to_remove, dtype=np.int) + to_remove_tracker = np.array(dets_to_remove, dtype=np.int32) # Apply preprocessing to remove all unwanted tracker dets. data['tracker_ids'][t] = np.delete(tracker_ids, to_remove_tracker, axis=0) @@ -444,14 +444,14 @@ def get_preprocessed_seq_data(self, raw_data, cls): gt_id_map[unique_gt_ids] = np.arange(len(unique_gt_ids)) for t in range(raw_data['num_timesteps']): if len(data['gt_ids'][t]) > 0: - data['gt_ids'][t] = gt_id_map[data['gt_ids'][t]].astype(np.int) + data['gt_ids'][t] = gt_id_map[data['gt_ids'][t]].astype(np.int32) if len(unique_tracker_ids) > 0: unique_tracker_ids = np.unique(unique_tracker_ids) tracker_id_map = np.nan * np.ones((np.max(unique_tracker_ids) + 1)) tracker_id_map[unique_tracker_ids] = np.arange(len(unique_tracker_ids)) for t in range(raw_data['num_timesteps']): if len(data['tracker_ids'][t]) > 0: - data['tracker_ids'][t] = tracker_id_map[data['tracker_ids'][t]].astype(np.int) + data['tracker_ids'][t] = tracker_id_map[data['tracker_ids'][t]].astype(np.int32) # Record overview statistics. data['num_tracker_dets'] = num_tracker_dets diff --git a/eval/posetrack21/posetrack21/trackeval/metrics/hota.py b/eval/posetrack21/posetrack21/trackeval/metrics/hota.py index 2700569..332597c 100644 --- a/eval/posetrack21/posetrack21/trackeval/metrics/hota.py +++ b/eval/posetrack21/posetrack21/trackeval/metrics/hota.py @@ -31,20 +31,20 @@ def eval_sequence(self, data): # Initialise results res = {} for field in self.float_array_fields + self.integer_array_fields: - res[field] = np.zeros((len(self.array_labels)), dtype=np.float) + res[field] = np.zeros((len(self.array_labels)), dtype=np.float32) for field in self.float_fields: res[field] = 0 # Return result quickly if tracker or gt sequence is empty if data['num_tracker_dets'] == 0: - res['HOTA_FN'] = data['num_gt_dets'] * np.ones((len(self.array_labels)), dtype=np.float) - res['LocA'] = np.ones((len(self.array_labels)), dtype=np.float) + res['HOTA_FN'] = data['num_gt_dets'] * np.ones((len(self.array_labels)), dtype=np.float32) + res['LocA'] = np.ones((len(self.array_labels)), dtype=np.float32) res['LocA(0)'] = 1.0 res = self._compute_final_fields(res) return res if data['num_gt_dets'] == 0: - res['HOTA_FP'] = data['num_tracker_dets'] * np.ones((len(self.array_labels)), dtype=np.float) - res['LocA'] = np.ones((len(self.array_labels)), dtype=np.float) + res['HOTA_FP'] = data['num_tracker_dets'] * np.ones((len(self.array_labels)), dtype=np.float32) + res['LocA'] = np.ones((len(self.array_labels)), dtype=np.float32) res['LocA(0)'] = 1.0 res = self._compute_final_fields(res) return res @@ -343,9 +343,9 @@ def _summary_result(self, metric, result): vals.append("{0:1.5g}".format(100 * np.mean(result))) # we have an array for each joint elif metric in self.integer_array_fields: - vals.append("{0:d}".format(np.mean(result).astype(np.int))) + vals.append("{0:d}".format(np.mean(result).astype(np.int32))) elif metric in self.integer_fields: - vals.append("{0:d}".format(result.astype(np.int))) + vals.append("{0:d}".format(result.astype(np.int32))) else: raise TrackEvalException(f"Unknown metric {metric}") diff --git a/eval/posetrack21/posetrack21/trackeval/metrics/hota_pose.py b/eval/posetrack21/posetrack21/trackeval/metrics/hota_pose.py index 65adac8..194dc7e 100644 --- a/eval/posetrack21/posetrack21/trackeval/metrics/hota_pose.py +++ b/eval/posetrack21/posetrack21/trackeval/metrics/hota_pose.py @@ -54,17 +54,17 @@ def eval_sequence(self, data): assert res['HOTA_TP'].shape[1] == 15 # Return result quickly if tracker or gt sequence is empty if data['num_tracker_dets'] == 0: - res['HOTA_FN'] = data['num_gt_joints'][None, :] * np.ones((len(self.array_labels), self.n_joints), dtype=np.float) - res['LocA'] = np.ones((len(self.array_labels), self.n_joints), dtype=np.float) - res['LocA(0)'] = np.ones((self.n_joints), dtype=np.float) + res['HOTA_FN'] = data['num_gt_joints'][None, :] * np.ones((len(self.array_labels), self.n_joints), dtype=np.float32) + res['LocA'] = np.ones((len(self.array_labels), self.n_joints), dtype=np.float32) + res['LocA(0)'] = np.ones((self.n_joints), dtype=np.float32) res = self._compute_final_fields(res, compute_avg=True) return res if data['num_gt_dets'] == 0: - res['HOTA_FP'] = data['num_tracker_joints'][None, :] * np.ones((len(self.array_labels), self.n_joints), dtype=np.float) - res['LocA'] = np.ones((len(self.array_labels), self.n_joints), dtype=np.float) - res['LocA(0)'] = np.ones((self.n_joints), dtype=np.float) + res['HOTA_FP'] = data['num_tracker_joints'][None, :] * np.ones((len(self.array_labels), self.n_joints), dtype=np.float32) + res['LocA'] = np.ones((len(self.array_labels), self.n_joints), dtype=np.float32) + res['LocA(0)'] = np.ones((self.n_joints), dtype=np.float32) res = self._compute_final_fields(res, compute_avg=True) return res @@ -330,10 +330,10 @@ def _summary_result(self, metric, result): # we have an array for each joint elif metric in self.integer_array_fields: for j in range(result.shape[1]): - vals.append("{0:d}".format(np.mean(result[:, j]).astype(np.int))) + vals.append("{0:d}".format(np.mean(result[:, j]).astype(np.int32))) elif metric in self.integer_fields: for j in range(result.shape[0]): - vals.append("{0:d}".format(result[j].astype(np.int))) + vals.append("{0:d}".format(result[j].astype(np.int32))) else: raise TrackEvalException(f"Unknown metric {metric}")