[Fix] remove tqdm & fix posec3d

demo/demo_posec3d.py

@@ -224,7 +224,8 @@ def main():
         modality='Pose',
         total_frames=num_frame)
     num_person = max([len(x) for x in pose_results])
-    num_keypoint = pose_results[0][0]['keypoints'].shape[0]
+    # Current PoseC3D models are trained on COCO-keypoints (17 keypoints)
+    num_keypoint = 17
     keypoint = np.zeros((num_person, num_frame, num_keypoint, 2),
                         dtype=np.float16)
     keypoint_score = np.zeros((num_person, num_frame, num_keypoint),

demo/demo_spatiotemporal_det.py

@@ -10,7 +10,6 @@
 import torch
 from mmcv import DictAction
 from mmcv.runner import load_checkpoint
-from tqdm import tqdm
 
 from mmaction.models import build_detector
 from mmaction.utils import import_module_error_func
@@ -225,11 +224,13 @@ def detection_inference(args, frame_paths):
                                           'trained on COCO')
     results = []
     print('Performing Human Detection for each frame')
-    for frame_path in tqdm(frame_paths):
+    prog_bar = mmcv.ProgressBar(len(frame_paths))
+    for frame_path in frame_paths:
         result = inference_detector(model, frame_path)
         # We only keep human detections with score larger than det_score_thr
         result = result[0][result[0][:, 4] >= args.det_score_thr]
         results.append(result)
+        prog_bar.update()
     return results
 
 
@@ -355,7 +356,9 @@ def main():
     predictions = []
 
     print('Performing SpatioTemporal Action Detection for each clip')
-    for timestamp, proposal in tqdm(zip(timestamps, human_detections)):
+    assert len(timestamps) == len(human_detections)
+    prog_bar = mmcv.ProgressBar(len(timestamps))
+    for timestamp, proposal in zip(timestamps, human_detections):
         if proposal.shape[0] == 0:
             predictions.append(None)
             continue
@@ -389,6 +392,7 @@ def main():
                         prediction[j].append((label_map[i + 1], result[i][j,
                                                                           4]))
             predictions.append(prediction)
+        prog_bar.update()
 
     results = []
     for human_detection, prediction in zip(human_detections, predictions):

setup.cfg

@@ -19,6 +19,6 @@ line_length = 79
 multi_line_output = 0
 known_standard_library = pkg_resources,setuptools
 known_first_party = mmaction
-known_third_party = cv2,decord,einops,joblib,matplotlib,mmcv,numpy,pandas,pytest,scipy,seaborn,titlecase,torch,tqdm,webcolors
+known_third_party = cv2,decord,einops,joblib,matplotlib,mmcv,numpy,pandas,pytest,scipy,seaborn,titlecase,torch,webcolors
 no_lines_before = STDLIB,LOCALFOLDER
 default_section = THIRDPARTY

tools/analysis/check_videos.py

@@ -4,9 +4,9 @@
 from functools import partial
 from multiprocessing import Manager, Pool, cpu_count
 
+import mmcv
 import numpy as np
 from mmcv import Config, DictAction
-from tqdm import tqdm
 
 from mmaction.datasets import PIPELINES, build_dataset
 
@@ -134,8 +134,9 @@ def _do_check_videos(lock, dataset, output_file, idx):
     ids = range(len(dataset))
 
     # start checking
-    for _ in tqdm(pool.imap_unordered(worker_fn, ids), total=len(ids)):
-        pass
+    prog_bar = mmcv.ProgressBar(len(dataset))
+    for _ in pool.imap_unordered(worker_fn, ids):
+        prog_bar.update()
     pool.close()
     pool.join()