shijies / dmmn Goto Github PK
View Code? Open in Web Editor NEWDeep Motion Modeling Network
Home Page: https://shijies.github.io/DMMN_Page/
Deep Motion Modeling Network
Home Page: https://shijies.github.io/DMMN_Page/
unable to find the folders for
"log_save_folder": "/media/ssm/data/dataset/amotd/test_logs/0808-67650/logs",
"image_save_folder": "/media/ssm/data/dataset/amotd/test_logs/0808-67650/images",
"weights_save_folder": "/media/ssm/data/dataset/amotd/test_logs/0808-67650/weights",
hi, can it be trained on the MOT17 dataset?
Facing difficulties while downloading the dataset from baidu outside of china. Do you have the dataset available on some other file hosting service?
Thank you
Hi,
I got an error when I tried to test the UA-DETRAC with the command:
python test_tracker_ua.py
The error message is as below:
Traceback (most recent call last):
File "test_tracker_ua.py", line 12, in
from layers.dmmn.tracker import Tracker, Config
File "/root/data/AIModel/DMMN/layers/dmmn/tracker.py", line 199, in
class TrackSet:
File "/root/data/AIModel/DMMN/layers/dmmn/tracker.py", line 215, in TrackSet
@staticmethod
File "/root/data/anaconda3/envs/py36/lib/python3.6/site-packages/deco/conc.py", line 44, in init
self.name = f.name
AttributeError: 'staticmethod' object has no attribute 'name'
Did I do anything wrong?
How could I fix it?
Thanks!
loading configure: config_gpu4_amot.json========
{
"dataset_name": "AMOTD",
"dataset_path": "~/data/omot_partial_dataset",
"phase": "train",
"frame_max_input_num": 16,
"frame_sample_scale": 2,
"parameter_frame_scale": 0.25,
"random_select_frame": false,
"min_valid_node_rate": 0.15,
"num_classes": 2,
"cuda": true,
"frame_size": 168,
"pixel_mean": [57, 52, 50],
"num_motion_model_param": 12,
"video_fps": 30.0,
"image_width": 1920,
"image_height": 1080,
"label_map": {
"vehicle": 1
},
"replace_map": {
"vehicle": 1
},
"test": {
"resume": "./test_logs/weights/ssdt67650.pth",
"dataset_type": "train",
"batch_size": 1,
"num_workers": 1,
"lr_decay_per_epoch": [1, 30, 45, 50],
"base_net_weights": null,
"log_save_folder": "./logs/test_logs/logs",
"image_save_folder": "./logs/test_logs/images",
"weights_save_folder": "./logs/test_logs/weights",
"sequence_list": "./dataset/amot/sequence_list_town02_train_part.txt",
"save_weight_per_epoch": 5,
"start_epoch": 0,
"end_epoch": 55,
"tensorboard": true,
"port": 6006,
"momentum": 0.9,
"weight_decay": 5e-4,
"gamma": 0.1,
"send_images": true,
"log_iters": true,
"run_mode": "debug",
"debug_save_image": false,
"debug_save_feature_map": false,
"save_track_data": true,
"contrast_lower": 0.5,
"contrast_upper": 1.5,
"saturation_lower": 0.5,
"saturation_upper": 1.5,
"hue_delta": 18.0,
"brightness_delta": 32,
"max_expand_ratio": 1.1,
"detect_bkg_label": 0,
"detect_top_k": 300,
"detect_conf_thresh": 0.3,
"detect_nms_thresh": 0.3,
"detect_exist_thresh": 0.5,
"tracker_min_iou_thresh": 0.001,
"tracker_min_visibility": 0.4
},
"train": {
"resume": null,
"batch_size": 8,
"num_workers": 0,
"learning_rate": 1e-3,
"lr_decay_per_epoch": [30, 50, 70, 90],
"base_net_weights": "./weights/resnext-101-64f-kinetics.pth",
"log_save_folder": "./logs/train_logs/log",
"image_save_folder": "./logs/train_logs/image",
"weights_save_folder": "./logs/train_logs/weights",
"sequence_list": "./dataset/amot/sequence_list_town02_train_part.txt",
"save_weight_per_epoch": 0.2,
"start_epoch": 0,
"end_epoch": 200,
"tensorboard": true,
"port": 6006,
"momentum": 0.9,
"weight_decay": 5e-4,
"gamma": 0.1,
"send_images": true,
"log_iters": true,
"run_mode": "release",
"debug_save_image": false,
"debug_save_feature_map": false,
"contrast_lower": 0.5,
"contrast_upper": 1.5,
"saturation_lower": 0.5,
"saturation_upper": 1.5,
"hue_delta": 18.0,
"brightness_delta": 32,
"max_expand_ratio": 1.1,
"static_possiblity": 0.05,
"loss_overlap_thresh": 0.5,
"loss_background_label": 0,
"dataset_overlap_thresh": 0.75
},
"frame_work":{
"temporal_dims": [8, 4, 2, 1, 1, 1],
"channel_dims": [256, 512, 1024, 2048, 2048, 2048],
"feature_maps": [42, 21, 11, 6, 3, 2],
"steps": [4, 8, 16, 28, 56, 84],
"min_sizes": [4, 16, 32, 64, 108, 146],
"max_sizes": [16, 32, 64, 108, 146, 176],
"aspect_ratios": [[1.5, 2], [2, 3], [2, 3], [2], [2], [2]],
"boxes_scales": [[0.8333, 0.6667, 0.5, 0.4], [0.8333, 0.5], [0.8333, 0.5], [0.5], [], []],
"variance": [0.1, 0.2],
"branch_cnn": 3,
"clip": true
},
"base_net":{
"mode": "feature",
"model_name": "resnext",
"model_depth": 101,
"resnet_shortcut": "B",
"arch": "resnext-101"
}
}
reading: ~/data/omot_partial_dataset/train/Town02/Clear/50/Easy_Camera_8.avi: 80%|███████████████████████▏ | 4/5 [00:11<00:02, 3.00s/it]
Loading base network...
Loading base net weights into state dict...
Finish
Timer: 8.9696 sec.
iter 0, 3078 || epoch: 0.0000 || Loss: 1485.7682 || Saving weights, iter: 0
Timer: 1.1191 sec.
iter 10, 3078 || epoch: 0.0032 || Loss: 747.8365 || Timer: 1.3092 sec.
iter 20, 3078 || epoch: 0.0065 || Loss: 543.7720 || Timer: 0.9543 sec.
iter 30, 3078 || epoch: 0.0097 || Loss: 483.3675 || Timer: 1.3921 sec.
iter 40, 3078 || epoch: 0.0130 || Loss: 365.0275 || Timer: 0.8044 sec.
iter 50, 3078 || epoch: 0.0162 || Loss: 326.6543 || Timer: 0.9312 sec.
iter 60, 3078 || epoch: 0.0195 || Loss: 366.3070 || Timer: 1.2947 sec.
iter 70, 3078 || epoch: 0.0227 || Loss: 369.2639 || Timer: 1.1285 sec.
iter 80, 3078 || epoch: 0.0260 || Loss: 350.6907 || Timer: 0.8387 sec.
iter 90, 3078 || epoch: 0.0292 || Loss: 279.8749 || Timer: 0.9261 sec.
iter 100, 3078 || epoch: 0.0325 || Loss: 373.5823 || Timer: 0.9241 sec.
iter 110, 3078 || epoch: 0.0357 || Loss: 280.2119 || Timer: 1.1693 sec.
iter 120, 3078 || epoch: 0.0390 || Loss: 273.2498 || Timer: 1.5808 sec.
iter 130, 3078 || epoch: 0.0422 || Loss: 292.5098 || Timer: 0.8698 sec.
iter 140, 3078 || epoch: 0.0455 || Loss: 318.0992 || Timer: 1.0063 sec.
iter 150, 3078 || epoch: 0.0487 || Loss: 252.3025 || Timer: 0.9630 sec.
iter 160, 3078 || epoch: 0.0520 || Loss: 265.3047 || Timer: 1.4840 sec.
iter 170, 3078 || epoch: 0.0552 || Loss: 302.5854 || Timer: 0.9548 sec.
iter 180, 3078 || epoch: 0.0585 || Loss: 285.4070 || Timer: 0.9854 sec.
iter 190, 3078 || epoch: 0.0617 || Loss: 364.6956 || Timer: 4.4852 sec.
iter 200, 3078 || epoch: 0.0650 || Loss: 330.2629 || Timer: 1.0476 sec.
iter 210, 3078 || epoch: 0.0682 || Loss: 230.9167 || Timer: 0.9508 sec.
iter 220, 3078 || epoch: 0.0715 || Loss: 244.9536 || Timer: 0.8630 sec.
iter 230, 3078 || epoch: 0.0747 || Loss: 251.5319 || Timer: 0.8497 sec.
iter 240, 3078 || epoch: 0.0780 || Loss: 263.2910 || Timer: 0.8375 sec.
iter 250, 3078 || epoch: 0.0812 || Loss: 249.8669 || Timer: 0.9519 sec.
iter 260, 3078 || epoch: 0.0845 || Loss: 240.7322 || Timer: 1.1921 sec.
iter 270, 3078 || epoch: 0.0877 || Loss: 263.4745 || Timer: 1.1276 sec.
iter 280, 3078 || epoch: 0.0910 || Loss: 231.0480 || Timer: 1.2120 sec.
iter 290, 3078 || epoch: 0.0942 || Loss: 319.0433 || Timer: 0.7737 sec.
iter 300, 3078 || epoch: 0.0975 || Loss: 253.2475 || Timer: 1.1360 sec.
iter 310, 3078 || epoch: 0.1007 || Loss: 197.9579 || Timer: 0.9226 sec.
iter 320, 3078 || epoch: 0.1040 || Loss: 269.4134 || Timer: 0.9375 sec.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.