- Implementation of Non-local Neural Block.
-
You can find different kinds of non-local block in lib/.
-
You can visualize the Non_local Attention Map by following the Running Steps shown below.
-
The code is tested on MNIST dataset. You can select the type of non-local block in lib/network.py.
-
If there is something wrong in my code, please contact me, thanks!
- python 3.7.7 (>=3.8)
- pytorch 1.4.0 (>=1.2.0; 2.0.0 works too)
- opencv 4.2.0.34 (others work too)
- In the first Non-local Layer.
- In the second Non-local Layer.
-
Select the type of non-local block in lib/network.py.
from lib.non_local_concatenation import NONLocalBlock2D from lib.non_local_gaussian import NONLocalBlock2D from lib.non_local_embedded_gaussian import NONLocalBlock2D from lib.non_local_dot_product import NONLocalBlock2D -
Run demo_MNIST_train.py with one GPU or multi GPU to train the Network. Then the weights will be save in weights/.
CUDA_VISIBLE_DEVICES=0,1 python demo_MNIST_train.py # Or train with Automatic Mixed Precision based on pytorch 1.6.0 CUDA_VISIBLE_DEVICES=0 python demo_MNIST_AMP_train_with_single_gpu.py -
Run nl_map_save.py to save NL_MAP of one test sample in nl_map_vis.
CUDA_VISIBLE_DEVICES=0,1 python nl_map_save.py -
Come into nl_map_vis/ and run nl_map_vis.py to visualize the NL_MAP. (tips: if the Non-local type you select is non_local_concatenation or non_local_dot_product (without Softmax operation), you may need to normalize NL_MAP in the visualize code)
python nl_map_save.py
-
Figure out how to implement the concatenation type, and add the code to lib/.
-
Fix the bug in lib/non_local.py (old version) when using multi-gpu. Someone shares the reason with me, and you can find it in here.
-
Fix the error of 3D pooling in lib/non_local.py (old version). Appreciate protein27 for pointing it out.
-
For convenience, I split the lib/non_local.py into four python files, and move the old versions (lib/non_loca.py and lib/non_local_simple_version.py) into lib/backup/.
-
Modify the code to support pytorch 0.4.1, and move the code supporting pytorch 0.3.1
to Non-Local_pytorch_0.3.1/. -
Test the code with pytorch 1.1.0 and it works.
-
Move the code supporting pytorch 0.4.1 and 1.1.0 to Non-Local_pytorch_0.4.1_to_1.1.0/ (In fact, I think it can also support pytorch 1.2.0).
-
In order to visualize NL_MAP, some code have been slightly modified. The code nl_map_save.py is added to save NL_MAP (two Non-local Layer) of one test sample. The code Non-local_pytorch/nl_map_vis.py is added to visualize NL_MAP. Besieds, the code is support pytorch 1.2.0.
-
The code also works well in pytorch 1.4.0.
-
The code also works well in pytorch 1.6.0. Add demo_MNIST_AMP_train_with_single_gpu.py with Automatic Mixed Precision Training (FP16), supported by pytorch 1.6.0. It can reduce GPU memory during training. What's more, if you use GPU 2080Ti (tensor cores), training speed can be increased. More details (such as how to train with multiple GPUs) can be found in here
-
Verify that the code works well in pytorch 1.7.0.
-
Verify that the code works well in pytorch 1.8.1.
-
Verify that the code works well in pytorch 1.9.0.
-
Verify that the code works well in pytorch 1.10.1.
-
Verify that the code works well in pytorch 1.11.0.
-
Verify that the code works well in pytorch 1.12.0.
-
Remove redundant the code
net.train()in training code files. -
Verify that the code works well in pytorch 1.13.0 and 2.0.0.
- Experiments on Charades dataset.
- Experiments on COCO dataset.
- Non-local ResNet-50 TSM (Paper) on Kinetics dataset. They report that their model achieves a good performance of 75.6% on Kinetics, which is even higher than Non-local ResNet-50 I3D (Here).