zhangzilongc / mmr Goto Github PK
View Code? Open in Web Editor NEWLicense: Apache License 2.0
License: Apache License 2.0
请问预训练模型mae_visualize_vit_base文件该放在文件夹的哪个位置
We sincerely look forward to an early release of the AeBAD dataset!
Hello, Thank you for your excellent work. I have a query regarding the anomaly map. In the paper, you mentioned the process of computing anomaly maps, but you used cosine similarity in the code. Is their anything I missing. Can you please elaborate on the difference between paper and implementation?
Code:
Line 74 in 22d1b17
您好,阅读完论文之后,没有特别理解用AeBAD-S数据集在训练及test时要怎么分配数据,想和您确认:
训练数据时:train中的所有图片都会作为输入,最终生成1个异常检测模型
test即推理时:按照same, background, illumination, 和view这4个维度分别推理,然后每个维度分别计算指标(如论文结果)吗?
期待您的回答,十分感谢!
I noticed that there isn't a license specified in the repository, would you be able to clarify if the project is intended for open-source sharing and collaboration?
If so, would you consider adding a specific license? Thank you very much.
跑aebad_v数据集时代码出现路径混乱问题
root@autodl-container-b8cd118952-a284ba93:~/autodl-tmp# sh AeBAD_V_run.sh
[06/06 10:11:25][INFO] main.py: 26: {'DATASET': {'domain_shift_category': 'same',
'imagesize': 224,
'name': 'aebad_V',
'resize': 256,
'subdatasets': ['AeBAD_V']},
'NUM_GPUS': 1,
'OUTPUT_DIR': './log_MMR_AeBAD_V_54',
'RNG_SEED': 54,
'TEST': {'VISUALIZE': CfgNode({'Random_sample': False, 'Sample_num': 40}),
'dataset_path': './AeBAD',
'enable': False,
'method': 'MMR',
'pixel_mode_verify': False,
'save_segmentation_images': False,
'save_video_segmentation_images': True},
'TEST_SETUPS': CfgNode({'batch_size': 32}),
'TRAIN': {'MMR': {'DA_low_limit': 0.7,
'DA_up_limit': 1.0,
'FPN_output_dim': (256, 512, 1024),
'feature_compression': False,
'finetune_mask_ratio': 0.4,
'layers_to_extract_from': ['layer1', 'layer2', 'layer3'],
'load_pretrain_model': True,
'model_chkpt': './mae_visualize_vit_base.pth',
'scale_factors': (4.0, 2.0, 1.0),
'test_mask_ratio': 0.0},
'backbone': 'wideresnet50',
'dataset_path': './AeBAD',
'enable': True,
'method': 'MMR',
'save_model': False},
'TRAIN_SETUPS': {'batch_size': 16,
'epochs': 200,
'learning_rate': 0.001,
'num_workers': 1,
'warmup_epochs': 50,
'weight_decay': 0.05}}
[06/06 10:11:25][INFO] main.py: 27: path_to_config is method_config/AeBAD_V/MMR.yaml
[06/06 10:11:25][INFO] main.py: 31: start training!
[06/06 10:11:25][INFO] train.py: 27: load dataset!
[06/06 10:11:25][INFO] train.py: 55: current individual_dataloader is aebad_V_AeBAD_V.
[06/06 10:11:25][INFO] train.py: 56: the data in current individual_dataloader aebad_V_AeBAD_V are 707.
[06/06 10:11:28][INFO] train.py: 73: train the decoder FPN of MMR from scratch!
[06/06 10:11:28][INFO] train.py: 76: MAE load meg: _IncompatibleKeys(missing_keys=['decoder_FPN_mask_token', 'decoder_FPN_pos_embed', 'simfp_2.0.weight', 'simfp_2.0.bias', 'simfp_2.1.weight', 'simfp_2.1.bias', 'simfp_2.3.weight', 'simfp_2.3.bias', 'simfp_2.4.weight', 'simfp_2.4.norm.weight', 'simfp_2.4.norm.bias', 'simfp_2.5.weight', 'simfp_2.5.norm.weight', 'simfp_2.5.norm.bias', 'simfp_3.0.weight', 'simfp_3.0.bias', 'simfp_3.1.weight', 'simfp_3.1.norm.weight', 'simfp_3.1.norm.bias', 'simfp_3.2.weight', 'simfp_3.2.norm.weight', 'simfp_3.2.norm.bias', 'simfp_4.0.weight', 'simfp_4.0.norm.weight', 'simfp_4.0.norm.bias', 'simfp_4.1.weight', 'simfp_4.1.norm.weight', 'simfp_4.1.norm.bias'], unexpected_keys=[])
Traceback (most recent call last):
File "main.py", line 47, in
main()
File "main.py", line 40, in main
train(cfg=cfg)
File "/root/autodl-tmp/tools/train.py", line 99, in train
MMR_instance.fit(individual_dataloader)
File "/root/autodl-tmp/models/MMR/MMR_pipeline.py", line 49, in fit
for image in individual_dataloader:
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 517, in next
data = self._next_data()
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1199, in _next_data
return self._process_data(data)
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1225, in _process_data
data.reraise()
File "/root/miniconda3/lib/python3.8/site-packages/torch/_utils.py", line 429, in reraise
raise self.exc_type(msg)
FileNotFoundError: Caught FileNotFoundError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 202, in _worker_loop
data = fetcher.fetch(index)
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/root/miniconda3/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 44, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/root/autodl-tmp/datasets/mvtec.py", line 89, in getitem
image = default_loader(image_path)
File "/root/miniconda3/lib/python3.8/site-packages/torchvision/datasets/folder.py", line 215, in default_loader
return pil_loader(path)
File "/root/miniconda3/lib/python3.8/site-packages/torchvision/datasets/folder.py", line 195, in pil_loader
with open(path, 'rb') as f:
FileNotFoundError: [Errno 2] No such file or directory: './AeBAD/AeBAD_V/train/good/./AeBAD/AeBAD_V/train/good/video1_train/28.jpg'
大佬可以帮忙解决一下吗
- [03/25 15:01:09][INFO] main.py: 27: path_to_config is method_config/MVTec/MMR.yaml
- [03/25 15:01:09][INFO] main.py: 31: start training!
- [03/25 15:01:09][INFO] train.py: 27: load dataset!
- [03/25 15:01:09][INFO] train.py: 55: current individual_dataloader is mvtec_01.
- [03/25 15:01:09][INFO] train.py: 56: the data in current individual_dataloader mvtec_01 are 400.
- [03/25 15:01:11][INFO] train.py: 73: train the decoder FPN of MMR from scratch!
- [03/25 15:01:11][INFO] train.py: 76: MAE load meg: _IncompatibleKeys(missing_keys=['decoder_FPN_mask_token', 'decoder_FPN_pos_embed', 'simfp_2.0.weight', 'simfp_2.0.bias', 'simfp_2.1.weight', 'simfp_2.1.bias', 'simfp_2.3.weight', 'simfp_2.3.bias', 'simfp_2.4.weight', 'simfp_2.4.norm.weight', 'simfp_2.4.norm.bias', 'simfp_2.5.weight', 'simfp_2.5.norm.weight', 'simfp_2.5.norm.bias', 'simfp_3.0.weight', 'simfp_3.0.bias', 'simfp_3.1.weight', 'simfp_3.1.norm.weight', 'simfp_3.1.norm.bias', 'simfp_3.2.weight', 'simfp_3.2.norm.weight', 'simfp_3.2.norm.bias', 'simfp_4.0.weight', 'simfp_4.0.norm.weight', 'simfp_4.0.norm.bias', 'simfp_4.1.weight', 'simfp_4.1.norm.weight', 'simfp_4.1.norm.bias'], unexpected_keys=[])
"Hello, although it can run, the log indicates that there are incompatible keys when loading the model parameters. Will this affect the experimental results, or can it be ignored? Looking forward to your reply."
Hello,
Thank you for your work.
I trained your model on the default setting using the Aebad_S dataset. But I cannot see the saved model although I use the save model flag as true.
Please elaborate on how to save the model and load it back to test on different images. How we can resume the model from ckpt. In addition, If possible how we can train on the custom dataset? Thank you for your time.
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.