search

WHU-Sigma / HyperSIGMA

The official repo for the paper "HyperSIGMA: Hyperspectral Intelligence Comprehension Foundation Model"
118 stars 10 forks source link

TypeError: spat_vit_b_rvsa() got an unexpected keyword argument 'args' #5

Closed kowo1001 closed 2 weeks ago

kowo1001 commented 1 month ago

Hello, I have two questions.

(1) I entered the command below, but I don't know why the error below occurs. $ bash train_complex.sh 1e-4 hypersigma 2 l2 100

====================Command Result Start====================== Namespace(prefix='hypersigma_gaussian', arch='hypersigma', batchSize=4, lr=0.0001, wd=0, loss='l2', testdir=None, sigma=None, training_dataset_path='./dataset/WDC/training/wdc.db', pretrain='/mnt/code/users/yuchunmiao/hypersigma-master/pre_train/checkpoint-400.pth', init='kn', no_cuda=False, from_scratch=False, pretrain_path='./pre_train/spat-vit-base-ultra-checkpoint-1599.pth', no_log=False, threads=1, seed=2018, resume=False, no_ropt=False, chop=False, resumePath=None, dataroot='/data/HSI_Data/ICVL64_31.db', clip=1000000.0, gpu_ids=[0], basedir='./output/original_hypersigma_1e-4_spat-vit-base-ultra-checkpoint-1599_batch4_warmup_l2_epoch_100_gaussian_new_fusion', epoch=100, update_lr=5e-05, meta_lr=5e-05, n_way=1, k_spt=2, k_qry=5, task_num=16, update_step=5, update_step_test=10) Cuda Acess: 1 => creating model 'hypersigma' load our vit fusion_new_v5 final models Traceback (most recent call last): File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_denoising_gaussian_wdc.py", line 22, in engine = Engine(opt) File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_setup.py", line 146, in init self.setup() File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_setup.py", line 173, in setup self.net = hypersigma(args=self.opt) TypeError: spat_vit_b_rvsa() got an unexpected keyword argument 'args' ====================Command Result End======================

Any help is highly appreciated ;)

Thanks!

and, (2) I made the Washington dc mall.tif file( dc.tif ) into a train_0.mat, train_1.mat file through utility/mat_data.py in Image Denoising. Then I'm trying to make a train dataset. How can I make two mat files into a wdc.db file?

Or is there any other way to create a wbc.db file in training_dataset_path?

Please, suggest me how to do this.

leonmakise commented 1 month ago

We have updated our code. Plz try it and let us know whether still exists some problems.

kowo1001 commented 1 month ago

When I took the recently updated source code and ran it again, the following error occurred. I don't know why the error below occurs.

Any help is highly appreciated 👍

Thank U :D

I entered the command below.

(venv) jwjang@gpusystem:~/project/hsi_foundation/HyperSIGMA/ImageDenoising$ bash train_complex.sh 1e-4 hypersigma 2 l2 100

====================Command Result Start====================== mkdir: cannot create directory ‘./output/original_hypersigma_1e-4_spat-vit-base-ultra-checkpoint-1599_batch4_warmup_l2_epoch_100_complex_s3_8point’: File exists Namespace(prefix='hypersigma_gaussian', arch='hypersigma', batchSize=4, lr=0.0001, wd=0, loss='l2', testdir=None, sigma=None, training_dataset_path='./dataset/WDC/training/wdc.db', pretrain='/mnt/code/users/yuchunmiao/hypersigma-master/pre_train/checkpoint-400.pth', init='kn', no_cuda=False, from_scratch=False, pretrain_path='./pre_train/spat-vit-base-ultra-checkpoint-1599.pth', no_log=False, threads=1, seed=2018, resume=False, no_ropt=False, chop=False, resumePath=None, dataroot='/data/HSI_Data/ICVL64_31.db', clip=1000000.0, gpu_ids=[0], basedir='./output/original_hypersigma_1e-4_spat-vit-base-ultra-checkpoint-1599_batch4_warmup_l2_epoch_100_complex_s3_8point', epoch=100, update_lr=5e-05, meta_lr=5e-05, n_way=1, k_spt=2, k_qry=5, task_num=16, update_step=5, update_step_test=10) Cuda Acess: 1 => creating model 'hypersigma' load our vit fusion_new_v5 final models Traceback (most recent call last): File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_denoising_complex_wdc.py", line 23, in engine = Engine(opt) File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_setup.py", line 146, in init self.setup() File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/hsi_setup.py", line 173, in setup self.net = hypersigma(args=self.opt) File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/models/hypersigma/model.py", line 142, in spat_vit_b_rvsa model = SSFusionFramework( File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/models/hypersigma/model.py", line 97, in init self.spat_encoder.init_weights('/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/pre_train/spat-vit-base-ultra-checkpoint-1599.pth') File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/models/hypersigma/Spatial.py", line 495, in init_weights checkpoint = torch.load(pretrained, map_location='cpu') File "/data/jwjang/project/hsi_project/denoising/HSDT/pythonenv/venv/lib/python3.10/site-packages/torch/serialization.py", line 1040, in load return _legacy_load(opened_file, map_location, pickle_module, pickle_load_args) File "/data/jwjang/project/hsi_project/denoising/HSDT/pythonenv/venv/lib/python3.10/site-packages/torch/serialization.py", line 1262, in _legacy_load magic_number = pickle_module.load(f, pickle_load_args) _pickle.UnpicklingError: invalid load key, '\xff'. ====================Command Result End======================

leonmakise commented 1 month ago

It seems to be error about the loaded pretrain model. May I know ur env details, especially torch version? :D

kowo1001 commented 1 month ago

It seems to be error about the loaded pretrain model. May I know ur env details, especially torch version? :D Thank U for ur reply :D

torch version: 2.3.1+cu121 python version: 3.10.12 and

Package Version

cafe 2.0.7 huggingface-hub 0.23.4 ipykernel 6.29.5 ipython 8.26.0 joblib 1.4.2 json5 0.9.25 jsonpatch 1.33 jsonpointer 3.0.0 jsonschema 4.22.0 jsonschema-specifications 2023.12.1 jupyter 1.0.0 jupyter_client 8.6.2 jupyter-console 6.6.3 jupyter_core 5.7.2 jupyter-events 0.10.0 jupyter-lsp 2.2.5 jupyter_server 2.14.1 jupyter_server_terminals 0.5.3 jupyterlab 4.2.3 jupyterlab_pygments 0.3.0 jupyterlab_server 2.27.2 jupyterlab_widgets 3.0.11 kiwisolver 1.4.5 lazy_loader 0.4 lmdb 1.5.1 matplotlib 3.9.0 matplotlib-inline 0.1.7 numpy 2.0.0 nvidia-cublas-cu12 12.1.3.1 nvidia-cuda-cupti-cu12 12.1.105 nvidia-cuda-nvrtc-cu12 12.1.105 nvidia-cuda-runtime-cu12 12.1.105 nvidia-cudnn-cu12 8.9.2.26 nvidia-cufft-cu12 11.0.2.54 nvidia-curand-cu12 10.3.2.106 nvidia-cusolver-cu12 11.4.5.107 nvidia-cusparse-cu12 12.1.0.106 nvidia-nccl-cu12 2.20.5 nvidia-nvjitlink-cu12 12.5.82 nvidia-nvtx-cu12 12.1.105 opencv-python 4.10.0.84 packaging 24.1 pandas 2.2.2 pillow 10.4.0 pip 24.1 protobuf 5.27.2 scikit-image 0.24.0 scikit-learn 1.5.1 scikit-plot 0.3.7 scipy 1.14.0 torch 2.3.1 torchlights 0.4.0 torchnet 0.0.4 torchstat 0.0.7 torchvision 0.18.1 urllib3 2.2.2 wheel 0.43.0

Do you need any other env details? I refer to the link below and it is said that there may be damage to the file itself. Is that correct?

https://wenku.csdn.net/answer/bd8127b290ba4812824b2d053cc5b58a

aledamo commented 1 month ago

I have the same issue when loading the pre-trained models, @leonmakise.

4

leonmakise commented 1 month ago

@kowo1001 @aledamo Got it. Seem to be problems with the models we uploaded to huggingface. Let our team do a quick check. Some other models have similar error logs like this. Plz stay tune on it and give us some time. Thanks~

leonmakise commented 1 month ago

We found files are inconsistent with SHA256 check. A new model has been uploaded to huggingface. Please have a try and tell us if it is right now. https://huggingface.co/WHU-Sigma/HyperSIGMA/blob/main/spat-vit-base-ultra-checkpoint-1599.pth @kowo1001 @aledamo

kowo1001 commented 1 month ago

We found files are inconsistent with SHA256 check. A new model has been uploaded to huggingface. Please have a try and tell us if it is right now. https://huggingface.co/WHU-Sigma/HyperSIGMA/blob/main/spat-vit-base-ultra-checkpoint-1599.pth @kowo1001 @aledamo

I tried again and the same error occurred. I think the same error occurs because of the problem below. Could you also change the spec-vit-base-ultra-checkpoint-1599.pth file and upload it back to huggingface? :D 👍

leonmakise commented 1 month ago

@kowo1001 Please check it now. https://huggingface.co/WHU-Sigma/HyperSIGMA/blob/main/spec-vit-base-ultra-checkpoint-1599.pth

kowo1001 commented 1 month ago

I think the first question mentioned earlier will be resolved. Thank you. 👍👍 However, there seems to be some issues regarding the second question.

Below is the content of the second question mentioned earlier. The input is a dc.tif file downloaded from the site below, and I created two mat files (train_0.mat, train_1.mat) using the create_WDC_dataset function of 'mat_data.py'. Then we tried to convert the two mat files (train_0.mat, train_1.mat) to the wdc.db file. I tried to create the data.mdb, lock.mdb, meta_info.txt files in the wdc.db folder through the createDCmall function of lmdb_data.py, but the meta_info.txt file was not created. Below is the related error.

https://engineering.purdue.edu/~biehl/MultiSpec/hyperspectral.html

(venv) jwjang@gpusystem:~/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility$ python lmdb_data.py

====================Command Result Start====================== create wdc... (926, 191, 64, 64) map size (GB): 6.4770263671874995 /data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility/WDC/wdc5.db Traceback (most recent call last): File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility/lmdb_data.py", line 95, in createDCmall() File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility/lmdb_data.py", line 85, in createDCmall create_lmdb_train( File "/data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility/lmdb_data.py", line 55, in create_lmdb_train env = lmdb.open(name+'.db', map_size=map_size, writemap=True) lmdb.InvalidParameterError: /data/jwjang/project/hsi_foundation/HyperSIGMA/ImageDenoising/utility/WDC/wdc.db: Invalid argument ====================Command Result End======================

@kowo1001 Please check it now. https://huggingface.co/WHU-Sigma/HyperSIGMA/blob/main/spec-vit-base-ultra-checkpoint-1599.pth

Thank U 👍

leonmakise commented 1 month ago

Hi @kowo1001 , we understand your concerns and have uploaded the training and testing data for you now. You can find them in https://huggingface.co/datasets/WHU-Sigma/HyperSIGMA_Datasets/tree/main/HyperSIGMA_denoising. Please have a try now.

kowo1001 commented 1 month ago

Hi @kowo1001 , we understand your concerns and have uploaded the training and testing data for you now. You can find them in https://huggingface.co/datasets/WHU-Sigma/HyperSIGMA_Datasets/tree/main/HyperSIGMA_denoising. Please have a try now.

Thank you for your kindness. I wonder how you made the training dataset and the test dataset. Could you explain it in detail or give me a link to refer to? Thanks 👍

miaoyuchun commented 1 month ago

Hi @kowo1001, we are unsure about the second question mentioned earlier. You can refer to the code released in SST, which our data processing pipeline fully follows.