ROMP icon indicating copy to clipboard operation
ROMP copied to clipboard

video test promblem

Open thinkingIsMagic opened this issue 4 years ago • 7 comments

Dear Arthur151, when test video on your method, last frame always gets mistake, hope you reply.

this is issue:

pygame 2.0.0 (SDL 2.0.12, python 3.7.9) Hello from the pygame community. https://www.pygame.org/contribute.html /home/neu307/miniconda3/envs/env_centerhmr/lib/python3.7/site-packages/quaternion/numba_wrapper.py:21: UserWarning:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Could not import from numba, which means that some parts of this code may run MUCH more slowly. You may wish to install numba. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

warnings.warn(warning_text) INFO - 2021-11-16 15:56:39,929 - acceleratesupport - No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate' INFO - 2021-11-16 15:56:40,399 - base - {'tab': 'hrnet_cm64_test', 'configs_yml': 'configs/video.yml', 'demo_image_folder': 'None', 'local_rank': 0, 'model_version': 1, 'multi_person': True, 'collision_aware_centermap': False, 'collision_factor': 0.2, 'kp3d_format': 'smpl24', 'eval': False, 'max_person': 64, 'input_size': 512, 'Rot_type': '6D', 'rot_dim': 6, 'centermap_conf_thresh': 0.25, 'centermap_size': 64, 'deconv_num': 0, 'model_precision': 'fp32', 'backbone': 'hrnet', 'gmodel_path': '../trained_models/ROMP_hrnet32.pkl', 'print_freq': 50, 'fine_tune': True, 'gpu': '0', 'batch_size': 64, 'val_batch_size': 1, 'nw': 4, 'calc_PVE_error': False, 'dataset_rootdir': '/home/neu307/wh/dataset/', 'high_resolution': True, 'save_best_folder': '/home/neu307/wh/checkpoints/', 'log_path': '/home/neu307/wh/log/', 'total_param_count': 85, 'smpl_mean_param_path': '/home/neu307/wh/ROMP-master/models/satistic_data/neutral_smpl_mean_params.h5', 'smpl_model': '/home/neu307/wh/ROMP-master/models/statistic_data/neutral_smpl_with_cocoplus_reg.txt', 'smplx_model': True, 'cam_dim': 3, 'beta_dim': 10, 'smpl_joint_num': 22, 'smpl_model_path': '/home/neu307/wh/ROMP-master/models', 'smpl_uvmap': '/home/neu307/wh/ROMP-master/models/smpl/uv_table.npy', 'smpl_female_texture': '/home/neu307/wh/ROMP-master/models/smpl/SMPL_sampleTex_f.jpg', 'smpl_male_texture': '/home/neu307/wh/ROMP-master/models/smpl/SMPL_sampleTex_m.jpg', 'smpl_J_reg_h37m_path': '/home/neu307/wh/ROMP-master/models/smpl/J_regressor_h36m.npy', 'smpl_J_reg_extra_path': '/home/neu307/wh/ROMP-master/models/smpl/J_regressor_extra.npy', 'kernel_sizes': [5], 'GPUS': 0, 'use_coordmaps': True, 'video_or_frame': True, 'input_video_path': '../demo/test_videos/C1/C1_single_2.mp4', 'webcam_mesh_color': 'LightCyan', 'save_mesh': False, 'save_centermap': False, 'save_dict_results': True, 'webcam': False, 'fps_save': 30, 'multiprocess': False} INFO - 2021-11-16 15:56:40,399 - base - ------------------------------------------------------------------ INFO - 2021-11-16 15:56:40,399 - base - start building model. Using ROMP v1 INFO - 2021-11-16 15:56:43,931 - train_utils - using fine_tune model: ../trained_models/ROMP_hrnet32.pkl INFO - 2021-11-16 15:56:44,085 - base - finished build model. Initialization finished! Running the code on video ../demo/test_videos/C1/C1_single_2.mp4 Processing video 0/164 Processing video 1/164 Processing video 2/164 Processing video 3/164 Processing video 4/164 Processing video 5/164 Processing video 6/164 Processing video 7/164 Processing video 8/164 Processing video 9/164 Processing video 10/164 Processing video 11/164 Processing video 12/164 Processing video 13/164 Processing video 14/164 Processing video 15/164 Processing video 16/164 Processing video 17/164 Processing video 18/164 Processing video 19/164 Processing video 20/164 Processing video 21/164 Processing video 22/164 Processing video 23/164 Processing video 24/164 Processing video 25/164 Processing video 26/164 Processing video 27/164 Processing video 28/164 Processing video 29/164 Processing video 30/164 Processing video 31/164 Processing video 32/164 Processing video 33/164 Processing video 34/164 Processing video 35/164 Processing video 36/164 Processing video 37/164 Processing video 38/164 Processing video 39/164 Processing video 40/164 Processing video 41/164 Processing video 42/164 Processing video 43/164 Processing video 44/164 Processing video 45/164 Processing video 46/164 Processing video 47/164 Processing video 48/164 Processing video 49/164 Processing video 50/164 Processing video 51/164 Processing video 52/164 Processing video 53/164 Processing video 54/164 Processing video 55/164 Processing video 56/164 Processing video 57/164 Processing video 58/164 Processing video 59/164 Processing video 60/164 Processing video 61/164 Processing video 62/164 Processing video 63/164 Processing video 64/164 Processing video 65/164 Processing video 66/164 Processing video 67/164 Processing video 68/164 Processing video 69/164 Processing video 70/164 Processing video 71/164 Processing video 72/164 Processing video 73/164 Processing video 74/164 Processing video 75/164 Processing video 76/164 Processing video 77/164 Processing video 78/164 Processing video 79/164 Processing video 80/164 Processing video 81/164 Processing video 82/164 Processing video 83/164 Processing video 84/164 Processing video 85/164 Processing video 86/164 Processing video 87/164 Processing video 88/164 Processing video 89/164 Processing video 90/164 Processing video 91/164 Processing video 92/164 Processing video 93/164 Processing video 94/164 Processing video 95/164 Processing video 96/164 Processing video 97/164 Processing video 98/164 Processing video 99/164 Processing video 100/164 Processing video 101/164 Processing video 102/164 Processing video 103/164 Processing video 104/164 Processing video 105/164 Processing video 106/164 Processing video 107/164 Processing video 108/164 Processing video 109/164 Processing video 110/164 Processing video 111/164 Processing video 112/164 Processing video 113/164 Processing video 114/164 Processing video 115/164 Processing video 116/164 Processing video 117/164 Processing video 118/164 Processing video 119/164 Processing video 120/164 Processing video 121/164 Processing video 122/164 Processing video 123/164 Processing video 124/164 Processing video 125/164 Processing video 126/164 Processing video 127/164 Processing video 128/164 Processing video 129/164 Processing video 130/164 Processing video 131/164 Processing video 132/164 Processing video 133/164 Processing video 134/164 Processing video 135/164 Processing video 136/164 Processing video 137/164 Processing video 138/164 Processing video 139/164 Processing video 140/164 Processing video 141/164 Processing video 142/164 Processing video 143/164 Processing video 144/164 Processing video 145/164 Processing video 146/164 Processing video 147/164 Processing video 148/164 Processing video 149/164 Processing video 150/164 Processing video 151/164 Processing video 152/164 Processing video 153/164 Processing video 154/164 Processing video 155/164 Processing video 156/164 Processing video 157/164 Processing video 158/164 Processing video 159/164 Processing video 160/164 Traceback (most recent call last): File "/home/neu307/wh/ROMP-master/src/core/test.py", line 235, in main() File "/home/neu307/wh/ROMP-master/src/core/test.py", line 224, in main demo.process_video(args.input_video_path) File "/home/neu307/wh/ROMP-master/src/core/test.py", line 114, in process_video outputs = self.single_image_forward(frame) File "/home/neu307/wh/ROMP-master/src/core/test.py", line 98, in single_image_forward meta_data = img_preprocess(image, '0', input_size=args.input_size, single_img_input=True) File "/home/neu307/wh/ROMP-master/src/core/../lib/models/../utils/demo_utils.py", line 31, in img_preprocess image = image[:,:,::-1] TypeError: 'NoneType' object is not subscriptable

Process finished with exit code 1

thinkingIsMagic avatar Nov 16 '21 08:11 thinkingIsMagic

Sorry about that. I guess you are running code on a computer with multiple GPUs. Please run code with "CUDA_VISIBLE_DEVICES=0 " For instance, when you run it in processing video, please run it via CUDA_VISIBLE_DEVICES=0 python -u -m romp.predict.video --configs_yml='configs/video.yml'

Arthur151 avatar Nov 16 '21 10:11 Arthur151

Sorry about that. I guess you are running code on a computer with multiple GPUs. Please run code with "CUDA_VISIBLE_DEVICES=0 " For instance, when you run it in processing video, please run it via CUDA_VISIBLE_DEVICES=0 python -u -m romp.predict.video --configs_yml='configs/video.yml'

屏幕快照 2021-11-16 下午7 28 38

i run it on pycharm, have setted the cuda=0

thinkingIsMagic avatar Nov 16 '21 11:11 thinkingIsMagic

Then that would be pretty weird. I am chasing for a deadline now. For now, you may use try except + repeat the last few frame to make it pass. After the deadline, I will look into it. Is this ok?

Arthur151 avatar Nov 16 '21 12:11 Arthur151

Then that would be pretty weird. I am chasing for a deadline now. For now, you may use try except + repeat the last few frame to make it pass. After the deadline, I will look into it. Is this ok?

thank you, i will try

thinkingIsMagic avatar Nov 16 '21 12:11 thinkingIsMagic

屏幕快照 2021-11-16 下午10 09 59

I add try-except, and it works. Thank u

thinkingIsMagic avatar Nov 16 '21 14:11 thinkingIsMagic

My appologize to the unpleasant bug. I will get back to this later! Please remind me of this if you are in a hurry.

Arthur151 avatar Nov 16 '21 14:11 Arthur151

@jiandandian2 I am back to this issue. I have just updated a new version of ROMP. I just test it with processing 161/162/163/164 frames. It works fine. Could you please try it when you are convenient?

Arthur151 avatar Dec 02 '21 04:12 Arthur151