Generative remote sensing
Environment
nohup jupyter-notebook --no-browser --port=8888 & https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh /hy-tmp/miniconda3 https://mirror.tuna.tsinghua.edu.cn/help/anaconda/ https://developer.aliyun.com/mirror/anaconda https://mirrors.sustech.edu.cn/help/anaconda.html#configuration https://mirrors.tuna.tsinghua.edu.cn/help/pypi/
conda config --set custom_channels.nvidia https://mirrors.cernet.edu.cn/anaconda-extra/cloud/ conda clean -i
ln -s /root/autodl-tmp/miniconda3 /root/miniconda3
gpushare-cli baidu down /2023_generative_remote_sensing/pre-trained/stylegan-xl/imagenet256.pkl .
git config --global user.name "yc-cui" git config --global user.email "2548846280@qq.com"
Datasets
gpushare-cli baidu down /2023_generative_remote_sensing/datasets/AID/AID_dataset.zip
unzip x -d ../decomp
unzip NASC_TG2.zip '*RGB/*' -d ../decomp/NASC_TG2Models
taming-transformers
gpushare-cli baidu down /2023_generative_remote_sensing/pre-trained/taming-transformers/vqgan_imagenet_f16_16384.ckpt /hy-tmp/2023_generative_remote_sensing/generative_models/taming-transformers/logs/
conda env create -f environment.yaml conda activate taming
curl -o 2021-04-03T19-39-50_cin_transformer.zip "https://app.koofr.net/content/links/90cbd5aa-ef70-4f5e-99bc-f12e5a89380e/files/get/2021-04-03T19-39-50_cin_transformer.zip?path=%2F&force"
python main.py --base configs/custom_cond_AID_test0.2.yaml -t True --resume logs/2023-05-08T22-57-48_custom_cond_AID_test0.2
python main.py --base configs/custom_vqgan_AID_test0.2.yaml -t True python main.py --base configs/custom_vqgan_AID_test0.2.yaml -t True ; /usr/bin/shutdown
python main.py --base configs/custom_transformer_AID_test0.2.yaml -t True --resume logs/2023-05-15T14-17-42_custom_transformer_AID_test0.2; /usr/bin/shutdown
42, 1126, 2000, 3407, 31415
ln -s /data/cyc/2023-generative-remote-sensing/datasets /data/cyc/2023-generative-remote-sensing/generative_models/taming-transformers/
ln -s /root/autodl-tmp/2023-generative-remote-sensing/datasets /root/autodl-tmp/2023-generative-remote-sensing/generative_models/taming-transformers/
ln -s /hy-tmp/2023-generative-remote-sensing/datasets /hy-tmp/2023-generative-remote-sensing/generative_models/taming-transformers/
ImportError: libSM.so.6: cannot open shared object file: No such file or directory apt-get update apt-get install libsm6
docker run --rm -d -v wandb:/vol -p 8080:8080 --name wandb-local wandb/local
VQGAN ImageNet (f=16), 16384 https://heibox.uni-heidelberg.de/d/a7530b09fed84f80a887/
python scripts/sample_fast.py -r logs/temp -n 50 -k 600 -t 1.0 -p 0.92 --batch_size 5 --classes 27,28,29
ln -s /data/cyc/2023-generative-remote-sensing/datasets /data/cyc/2023-generative-remote-sensing/generative_models/stylegan-xl/
python dataset_tool.py --source=./datasets/data/cls/AID_test0.2/train.flist --dest=./data/AID_test0.2_train_256.zip --resolution=256x256 --transform=center-crop
python dataset_tool.py --source=./datasets/data/cls/AID_test0.2/train.flist --dest=./data/AID_test0.2_train_128.zip --resolution=128x128 --transform=center-crop
CUDA_VISIBLE_DEVICES=1 python train.py --outdir=./training-runs/AID_test0.2_train_256 --cfg=stylegan3-t --data=./data/AID_test0.2_train_256.zip --gpus=1 --batch=4 --mirror=1 --snap 10 --batch-gpu 4 --kimg 100 --syn_layers 31 --cond True --seed=42 --resume ./logs/imagenet256.pkl
fatal error: cuda_runtime_api.h: No such file or directory HawkAaron/warp-transducer#15 (comment) locate cuda_runtime_api.h export CUDA_HOME=/usr/local/cuda-11.4 # change to your path export CUDA_TOOLKIT_ROOT_DIR=$CUDA_HOME export LD_LIBRARY_PATH="$CUDA_HOME/extras/CUPTI/lib64:$LD_LIBRARY_PATH" export LIBRARY_PATH=$CUDA_HOME/lib64:$LIBRARY_PATH export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH export CFLAGS="-I$CUDA_HOME/include $CFLAGS"
Downloading: "https://dl.fbaipublicfiles.com/deit/deit_small_distilled_patch16_224-649709d9.pth" to /home/dell/.cache/torch/hub/checkpoints/deit_small_distilled_patch16_224-649709d9.pth wget -O ~/.cache/torch/hub/checkpoints/deit_small_distilled_patch16_224-649709d9.pth "https://dl.fbaipublicfiles.com/deit/deit_small_distilled_patch16_224-649709d9.pth"
RuntimeError: No such operator aten::cudnn_convolution_backward_weight autonomousvision/stylegan-xl#23 (comment) https://discuss.pytorch.org/t/at-has-no-member-cudnn-convolution-xxx/163716/4 nv-tlabs/GET3D#96 (comment)
OK: https://github.com/NVlabs/stylegan3/commit/407db86e6fe432540a22515310188288687858fa
mamba env create -f environment.yml
ln -s /data/cyc/2023-generative-remote-sensing/datasets /data/cyc/2023-generative-remote-sensing/generative_models/guided-diffusion/
ln -s /root/autodl-tmp/2023-generative-remote-sensing/datasets /root/autodl-tmp/2023-generative-remote-sensing/generative_models/guided-diffusion/
bash scripts/image_train.sh configs/custom_train_AID_test0.2.yaml bash scripts/image_sample.sh configs/custom_sample_AID_test0.2.yaml
script_util.py NUM_CLASSES = 30 dist_util.py CUDA_VISIBLE_DEVICES image_datasets.py num_workers=14
zip -r AID_test0.2.zip AID_test0.2/* tar czvf AID_test0.tar.gz AID_test0.2/*