Bump Version to OpenSTL V0.3.0

README.md

@@ -22,6 +22,7 @@
 [📘Documentation](https://openstl.readthedocs.io/en/latest/) |
 [🛠️Installation](docs/en/install.md) |
 [🚀Model Zoo](docs/en/model_zoos/video_benchmarks.md) |
+[👀Visualization](docs/en/visualization/video_visualization.md) |
 [🆕News](docs/en/changelog.md)
 
 ## Introduction
@@ -65,7 +66,7 @@ OpenSTL is a comprehensive benchmark for spatio-temporal predictive learning, en
 
 ## News and Updates
 
-[2023-05-28] `OpenSTL` v0.2.0 is released and enhanced. Some bugs in the training loop, datasets, and dataloaders are fixed.
+[2023-06-19] `OpenSTL` v0.3.0 is released and will be enhanced in [#25](https://github.com/chengtan9907/OpenSTL/issues/25).
 
 ## Installation
 

configs/weather/mv_4_s6_5_625/ConvLSTM.py

@@ -0,0 +1,23 @@
+method = 'ConvLSTM'
+# reverse scheduled sampling
+reverse_scheduled_sampling = 0
+r_sampling_step_1 = 25000
+r_sampling_step_2 = 50000
+r_exp_alpha = 5000
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '128,128,128,128'
+filter_size = 5
+stride = 1
+patch_size = 2
+layer_norm = 0
+
+# training
+lr = 5e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/MAU.py

@@ -0,0 +1,22 @@
+method = 'MAU'
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '64,64,64,64'
+filter_size = 5
+stride = 1
+patch_size = 1
+layer_norm = 0
+sr_size = 4
+tau = 5
+cell_mode = 'normal'
+model_mode = 'normal'
+
+# training
+lr = 5e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/MIM.py

@@ -0,0 +1,25 @@
+method = 'MIM'
+# reverse scheduled sampling
+reverse_scheduled_sampling = 0
+r_sampling_step_1 = 25000
+r_sampling_step_2 = 50000
+r_exp_alpha = 5000
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '128,128,128,128'
+filter_size = 5
+stride = 1
+patch_size = 4
+layer_norm = 0
+
+# training
+lr = 1e-4
+batch_size = 16
+val_batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0
+drop_last = True

configs/weather/mv_4_s6_5_625/PhyDNet.py

@@ -0,0 +1,9 @@
+method = 'PhyDNet'
+# model
+patch_size = 2
+
+# training
+lr = 5e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/PredRNN.py

@@ -0,0 +1,23 @@
+method = 'PredRNN'
+# reverse scheduled sampling
+reverse_scheduled_sampling = 0
+r_sampling_step_1 = 25000
+r_sampling_step_2 = 50000
+r_exp_alpha = 5000
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '128,128,128,128'
+filter_size = 5
+stride = 1
+patch_size = 2
+layer_norm = 0
+
+# training
+lr = 1e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/PredRNNpp.py

@@ -0,0 +1,23 @@
+method = 'PredRNNpp'
+# reverse scheduled sampling
+reverse_scheduled_sampling = 0
+r_sampling_step_1 = 25000
+r_sampling_step_2 = 50000
+r_exp_alpha = 5000
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '128,128,128,128'
+filter_size = 5
+stride = 1
+patch_size = 2
+layer_norm = 0
+
+# training
+lr = 1e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/PredRNNv2.py

@@ -0,0 +1,24 @@
+method = 'PredRNNv2'
+# reverse scheduled sampling
+reverse_scheduled_sampling = 1
+r_sampling_step_1 = 25000
+r_sampling_step_2 = 50000
+r_exp_alpha = 5000
+# scheduled sampling
+scheduled_sampling = 1
+sampling_stop_iter = 50000
+sampling_start_value = 1.0
+sampling_changing_rate = 0.00002
+# model
+num_hidden = '128,128,128,128'
+filter_size = 5
+stride = 1
+patch_size = 2
+layer_norm = 0
+decouple_beta = 0.1
+
+# training
+lr = 5e-4
+batch_size = 16
+sched = 'cosine'
+warmup_epoch = 0

configs/weather/mv_4_s6_5_625/SimVP_HorNet.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_IncepU.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_MLPMixer.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_MogaNet.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_Poolformer.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 1e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_Swin.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_Uniformer.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 1e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_VAN.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 5e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_ViT.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 1e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/SimVP_gSTA.py

@@ -9,7 +9,7 @@
 N_S = 2
 
 # training
-# lr = 1e-3
+lr = 1e-3
 batch_size = 16
 drop_path = 0.1
 sched = 'cosine'

configs/weather/mv_4_s6_5_625/TAU.py

@@ -0,0 +1,16 @@
+method = 'TAU'
+# model
+spatio_kernel_enc = 3
+spatio_kernel_dec = 3
+model_type = 'tau'
+hid_S = 32
+hid_T = 256
+N_T = 8
+N_S = 2
+alpha = 0.1
+# training
+lr = 5e-3
+batch_size = 16
+drop_path = 0.1
+sched = 'cosine'
+warmup_epoch = 5

docs/en/changelog.md

@@ -1,5 +1,17 @@
 ## Changelog
 
+### v0.3.0 (19/06/2023)
+
+Release version to OpenSTL V0.3.0 as [#25](https://github.com/chengtan9907/OpenSTL/issues/25).
+
+#### New Features
+
+* Support visualization tools in [vis_video](https://github.com/chengtan9907/OpenSTL/tree/master/tools/visualizations/vis_video.py), config files in [configs](https://github.com/chengtan9907/OpenSTL/tree/master/configs), and trained files (models, logs, and visualizations) in [v0.3.0](https://github.com/chengtan9907/OpenSTL/releases/tag/v0.3.0) of STL methods on various datasets (on updating).
+
+#### Update Documents
+
+* Update benchmark results of video prediction, traffic prediction, and weather prediction benchmarks in `docs/en/model_zoos`.
+
 ### v0.2.0 (21/04/2023)
 
 Release version to OpenSTL V0.2.0 as [#20](https://github.com/chengtan9907/OpenSTL/issues/20).
@@ -33,7 +45,7 @@ Release version to OpenSTL V0.2.0 as [#20](https://github.com/chengtan9907/OpenS
 * Fig bugs of building distributed dataloaders and preparation of DDP training.
 * Fix bugs of some STL methods (CrevNet, DMVFN, PreDNet, and TAU).
 * Fix bugs in datasets: fixing Caltech dataset for evaluation (28/05/2023 updating [Baidu Cloud](https://pan.baidu.com/s/1fudsBHyrf3nbt-7d42YWWg?pwd=kjfk)).
-* Fix the bug of `PSNR` (changing the implementation from E3D-LSTM to the corrent version) and update results in the benchmarks.
+* Fix the bug of `PSNR` (changing the implementation from E3D-LSTM to the current version) and update results in the benchmarks.
 
 ### v0.1.0 (18/02/2023)
 

docs/en/model_zoos/video_benchmarks.md

@@ -302,9 +302,9 @@ For a fair comparison of different methods, we report the best results when mode
 | E3D-LSTM     |  50 epoch |  4xbs4 |  60.9M |  542.0 |   7 | 143.3 | 1442.5 | 0.9803 | 32.52 | 0.04133 | model \| log |
 | PredNet      |  50 epoch | 1xbs16 |  12.5M |   13.7 | 176 |  |  |  |  |  | model \| log |
 | PhyDNet      |  50 epoch | 1xbs16 |   4.2M |   19.1 |  57 |  |  |  |  |  | model \| log |
-| MAU          |  50 epoch | 1xbs16 |  20.2M |  105.0 |   6 |  |  |  |  |  | model \| log |
+| MAU          |  50 epoch | 1xbs16 |  20.2M |  105.0 |   6 | 127.3 | 1577.0 | 0.9812 | 33.33 | 0.03561 | model \| log |
 | MIM          |  50 epoch |  4xbs4 |  47.6M | 1051.0 |  17 | 112.1 | 1467.1 | 0.9829 | 33.97 | 0.03338 | model \| log |
-| PredRNN      |  50 epoch | 1xbs16 |  24.6M |  704.0 |  25 | 113.2 | 1458.3 | 0.9831 | 33.94 | 0.032450 | model \| log |
+| PredRNN      |  50 epoch | 1xbs16 |  24.6M |  704.0 |  25 | 113.2 | 1458.3 | 0.9831 | 33.94 | 0.03245 | model \| log |
 | PredRNN++    |  50 epoch | 1xbs16 |  39.3M | 1033.0 |  18 | 110.0 | 1452.2 | 0.9832 | 34.02 | 0.03196 | model \| log |
 | PredRNN.V2   |  50 epoch | 1xbs16 |  24.6M |  708.0 |  24 | 114.9 | 1484.7 | 0.9827 | 33.84 | 0.03334 | model \| log |
 | DMVFN        |  50 epoch | 1xbs16 |   8.6M |   63.6 | 341 | 109.3 | 1449.3 | 0.9833 | 34.05 | 0.03189 | model \| log |
@@ -321,7 +321,7 @@ Since the hidden Translator in [SimVP](https://arxiv.org/abs/2211.12509) can be
 | IncepU (SimVPv1) |  50 epoch | 1xbs16 |  41.2M | 197.0 |  26 | 115.8 | 1511.5 | 0.9822 | 33.73 | 0.03467 | model \| log |
 | gSTA (SimVPv2)   |  50 epoch | 1xbs16 |  11.3M |  74.6 |  52 | 108.4 | 1441.0 | 0.9834 | 34.08 | 0.03224 | model \| log |
 | ViT              |  50 epoch |  4xbs4 |  28.3M | 239.0 |  17 | 136.3 | 1603.5 | 0.9796 | 33.10 | 0.03729 | model \| log |
-| Swin Transformer |  50 epoch | 1xbs16 |  38.8M | 188.0 |  28 |  |  |  |  |  | model \| log |
+| Swin Transformer |  50 epoch | 1xbs16 |  38.8M | 188.0 |  28 | 133.2 | 1599.7 | 0.9799 | 33.16 | 0.03766 | model \| log |
 | Uniformer        |  50 epoch |  4xbs4 |  27.7M | 211.0 |  14 | 116.3 | 1497.7 | 0.9824 | 33.76 | 0.03385 | model \| log |
 | MLP-Mixer        |  50 epoch | 1xbs16 |  47.0M | 164.0 |  34 | 125.7 | 1511.9 | 0.9819 | 33.49 | 0.03417 | model \| log |
 | ConvMixer        |  50 epoch | 1xbs16 |   3.1M |  39.4 |  84 | 115.8 | 1527.4 | 0.9822 | 33.67 | 0.03436 | model \| log |