Hi, would you be interested in adding MultiMAE to Hugging Face? The Hub offers free hosting, and it would make your work more accessible and visible to the rest of the ML community. Models/datasets/spaces(web demos) can be added to a user account or organization similar to github.

Example from other organizations: Keras: https://huggingface.co/keras-io Microsoft: https://huggingface.co/microsoft Facebook: https://huggingface.co/facebook

Example spaces with repos: github: https://github.com/salesforce/BLIP Spaces: https://huggingface.co/spaces/salesforce/BLIP

github: https://github.com/facebookresearch/omnivore Spaces: https://huggingface.co/spaces/akhaliq/omnivore

and here are guides for adding spaces/models/datasets to your org

How to add a Space: https://huggingface.co/blog/gradio-spaces how to add models: https://huggingface.co/docs/hub/adding-a-model uploading a dataset: https://huggingface.co/docs/datasets/upload_dataset.html

Please let us know if you would be interested and if you have any questions, we can also help with the technical implementation.

Hey, Thank you for providing the code for the paper. The paper is really interesting and the project page is very well done!

I was wondering whether you've tested the performance of linear probing on the RGB image when trained with all 3 modalities. The results of the original MAE paper were not very good, it is interesting to understand if the additional supervision creates better representations that translate into better linear probing scores.

Thanks, Eliahu

Thank you for your great Multi-MAE, We observed that in https://github.com/EPFL-VILAB/MultiMAE/blob/main/utils/datasets.py line 96, you use img = torch.Tensor(np.array(task_dict[task]) / 2 ** 16). Can you tell me the reason why depth map should be divided 2**16 ? Is there any problems without this operation?

Hi, I have successful made the pesudo labels and trained ‘rgb’ in/out-domain multimae model.

But when I trained model with 'rgb-semseg' in/out-domain, I met an error in multimae/input_adapters.py line 232

# Create patches [B, C, H, W] -> [B, (H*W), C]
x_patch = rearrange(self.proj(x), 'b d nh nw -> b (nh nw) d')

The full log is log.txt. x.size() is [batchsize, 64, 56, 56] before line 232. I can't find out what's wrong.

What's more, I don't know why the pseudo semeg label image resize into 1/4 (that is 224*224->56*56)in utils/datasets.py line 105

# Convert to Tensor
for task in task_dict:
    if task in ['depth']:
        img = torch.Tensor(np.array(task_dict[task]) / 2 ** 16)
        img = img.unsqueeze(0)  # 1 x H x W
    elif task in ['rgb']:
        img = TF.to_tensor(task_dict[task])
        img = TF.normalize(img, mean=self.rgb_mean, std=self.rgb_std)
    elif task in ['semseg', 'semseg_coco']:
        # TODO: add this to a config instead
        # Rescale to 0.25x size (stride 4)
        scale_factor = 0.25
        img = task_dict[task].resize((int(self.input_size * scale_factor), int(self.input_size * scale_factor)))
        # Using pil_to_tensor keeps it in uint8, to_tensor converts it to float (rescaled to [0, 1])
        img = TF.pil_to_tensor(img).to(torch.long).squeeze(0)

and then use nn.Conv2d in multimae/input_adapters.py line 198

if self.interpolate_class_emb:
    self.proj = nn.Sequential(
        nn.Upsample(scale_factor=(1 / self.P_H, 1 / self.P_W),
                    mode='bilinear'),  # Actually a downsample operation
        nn.Conv2d(in_channels=self.dim_class_emb, out_channels=self.dim_tokens,
                    kernel_size=1, stride=1),
    )
else:
    self.proj = nn.Conv2d(
        in_channels=self.dim_class_emb, out_channels=self.dim_tokens,
        kernel_size=(self.P_H, self.P_W), stride=(self.P_H, self.P_W)
    )
)

Thank you for any help.

Hey awesome work! I am trying to figure out how to modify the demo notebook to use the regular MAE instead of multiMAE. In particular i comment out all depth and semseg info but the resulting image infilling looks corrupted. Could you by chance share an example of proper usage of the regular MAE weights? Thanks so much for the help!

Hi, I am pseudo-tagging the imagenet-1k, and encountering some difficulties.

Firstly, I wonder what would happen if the classes of semeg are more than 255? How to use one channel depth png image to represent them? (Although COCO datasets is only 80 classes, the imagenet is more than 255 classes when fine-tuning)

Secondly, on the example of Colab notebook, the rgb2depth model of DPT could not input any size of imagenet pictures. How could we save all the pseudo labels down before the data augmentation cutting it into 224*224? We need to align the original images with the pseudo labeled image should we?

Thank you for any help.

Hi,

I am trying to pretrain the code on Celeb-HQ dataset and I sucessfully created respective grayscale depth maps(PNG) and grayscale segmentation(PNG) for pretraining. However, when i try to train "OMP_NUM_THREADS=1 torchrun --nproc_per_node=8 run_pretraining_multimae.py --config cfgs/pretrain/multimae-b_98_rgb+-depth-semseg_1600e.yaml --data_path /home/gargatik/gargatik/Datasets/copy/multimae/train"

I am facing the issue:

Start_______________________

../aten/src/ATen/native/cuda/Indexing.cu:975: indexSelectLargeIndex: block: [92,0,0], thread: [30,0,0] Assertion srcIndex < srcSelectDimSize failed. ../aten/src/ATen/native/cuda/Indexing.cu:975: indexSelectLargeIndex: block: [92,0,0], thread: [31,0,0] Assertion srcIndex < srcSelectDimSize failed. Traceback (most recent call last): File "run_pretraining_multimae.py", line 585, in main(opts) File "run_pretraining_multimae.py", line 414, in main train_stats = train_one_epoch( File "run_pretraining_multimae.py", line 501, in train_one_epoch preds, masks = model( File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 1008, in forward output = self._run_ddp_forward(*inputs, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 969, in _run_ddp_forward return module_to_run(*inputs[0], **kwargs[0]) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/multimae.py", line 312, in forward input_task_tokens = { File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/multimae.py", line 313, in domain: self.input_adaptersdomain File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/input_adapters.py", line 232, in forward x_patch = rearrange(self.proj(x), 'b d nh nw -> b (nh nw) d') File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, *kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 457, in forward return self._conv_forward(input, self.weight, self.bias) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 453, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR terminate called after throwing an instance of 'c10::CUDAError' what(): CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Exception raised from createEvent at ../aten/src/ATen/cuda/CUDAEvent.h:166 (most recent call first): frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x3e (0x7f1c685031ee in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libc10.so) frame #1: + 0xf3c2d (0x7f1caad91c2d in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so) frame #2: + 0xf6f6e (0x7f1caad94f6e in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so) frame #3: + 0x463418 (0x7f1cba0f6418 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #4: c10::TensorImpl::release_resources() + 0x175 (0x7f1c684ea7a5 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libc10.so) frame #5: + 0x35f2f5 (0x7f1cb9ff22f5 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #6: + 0x679288 (0x7f1cba30c288 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #7: THPVariable_subclass_dealloc(_object) + 0x2d5 (0x7f1cba30c655 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #8: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5ccad3] frame #9: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5d270c] frame #10: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5ec780] frame #11: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5441f8] frame #12: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x54424a] frame #13: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x54424a] frame #14: PyDict_SetItemString + 0x536 (0x5d1686 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #15: PyImport_Cleanup + 0x79 (0x684619 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #16: Py_FinalizeEx + 0x7f (0x67f8af in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #17: Py_RunMain + 0x32d (0x6b70fd in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #18: Py_BytesMain + 0x2d (0x6b736d in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #19: __libc_start_main + 0xf3 (0x7f1cd8fc10b3 in /lib/x86_64-linux-gnu/libc.so.6) frame #20: _start + 0x2e (0x5fa5ce in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python)

Traceback (most recent call last): File "run_pretraining_multimae.py", line 585, in main(opts) File "run_pretraining_multimae.py", line 414, in main train_stats = train_one_epoch( File "run_pretraining_multimae.py", line 501, in train_one_epoch preds, masks = model( File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 1008, in forward output = self._run_ddp_forward(*inputs, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 969, in _run_ddp_forward return module_to_run(*inputs[0], **kwargs[0]) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/multimae.py", line 312, in forward input_task_tokens = { File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/multimae.py", line 313, in domain: self.input_adaptersdomain File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/inpaint_proj/MultiMAE/multimae/input_adapters.py", line 232, in forward x_patch = rearrange(self.proj(x), 'b d nh nw -> b (nh nw) d') File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 457, in forward return self._conv_forward(input, self.weight, self.bias) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 453, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR You can try to repro this exception using the following code snippet. If that doesn't trigger the error, please include your original repro script when reporting this issue.

import torch torch.backends.cuda.matmul.allow_tf32 = False torch.backends.cudnn.benchmark = True torch.backends.cudnn.deterministic = False torch.backends.cudnn.allow_tf32 = True data = torch.randn([256, 64, 56, 56], dtype=torch.half, device='cuda', requires_grad=True).to(memory_format=torch.channels_last) net = torch.nn.Conv2d(64, 768, kernel_size=[4, 4], padding=[0, 0], stride=[4, 4], dilation=[1, 1], groups=1) net = net.cuda().half().to(memory_format=torch.channels_last) out = net(data) out.backward(torch.randn_like(out)) torch.cuda.synchronize()

ConvolutionParams data_type = CUDNN_DATA_HALF padding = [0, 0, 0] stride = [4, 4, 0] dilation = [1, 1, 0] groups = 1 deterministic = false allow_tf32 = true input: TensorDescriptor 0xc853ff10 type = CUDNN_DATA_HALF nbDims = 4 dimA = 256, 64, 56, 56, strideA = 200704, 1, 3584, 64, output: TensorDescriptor 0xc8540270 type = CUDNN_DATA_HALF nbDims = 4 dimA = 256, 768, 14, 14, strideA = 150528, 1, 10752, 768, weight: FilterDescriptor 0x819c34f0 type = CUDNN_DATA_HALF tensor_format = CUDNN_TENSOR_NHWC nbDims = 4 dimA = 768, 64, 4, 4, Pointer addresses: input: 0x7f12aa000000 output: 0x7f12ca000000 weight: 0x7f13d9200c00

terminate called after throwing an instance of 'c10::CUDAError' what(): CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Exception raised from createEvent at ../aten/src/ATen/cuda/CUDAEvent.h:166 (most recent call first): frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x3e (0x7f147c2811ee in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libc10.so) frame #1: + 0xf3c2d (0x7f14beb0fc2d in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so) frame #2: + 0xf6f6e (0x7f14beb12f6e in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so) frame #3: + 0x463418 (0x7f14cde74418 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #4: c10::TensorImpl::release_resources() + 0x175 (0x7f147c2687a5 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libc10.so) frame #5: + 0x35f2f5 (0x7f14cdd702f5 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #6: + 0x679288 (0x7f14ce08a288 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #7: THPVariable_subclass_dealloc(_object*) + 0x2d5 (0x7f14ce08a655 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/lib/libtorch_python.so) frame #8: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5ccad3] frame #9: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5d270c] frame #10: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5ec780] frame #11: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x5441f8] frame #12: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x54424a] frame #13: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python() [0x54424a] frame #14: PyDict_SetItemString + 0x536 (0x5d1686 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #15: PyImport_Cleanup + 0x79 (0x684619 in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #16: Py_FinalizeEx + 0x7f (0x67f8af in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #17: Py_RunMain + 0x32d (0x6b70fd in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #18: Py_BytesMain + 0x2d (0x6b736d in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python) frame #19: __libc_start_main + 0xf3 (0x7f14ecd3f0b3 in /lib/x86_64-linux-gnu/libc.so.6) frame #20: _start + 0x2e (0x5fa5ce in /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python)

WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182198 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182199 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182200 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182202 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182203 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182204 closing signal SIGTERM WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 182205 closing signal SIGTERM ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: -6) local_rank: 3 (pid: 182201) of binary: /mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/python Traceback (most recent call last): File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/bin/torchrun", line 8, in sys.exit(main()) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/distributed/elastic/multiprocessing/errors/init.py", line 345, in wrapper return f(*args, **kwargs) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/distributed/run.py", line 761, in main run(args) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/distributed/run.py", line 752, in run elastic_launch( File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 131, in call return launch_agent(self._config, self._entrypoint, list(args)) File "/mnt/train-data-3-ssd/gargatik/virtual_env/multimae/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 245, in launch_agent raise ChildFailedError( torch.distributed.elastic.multiprocessing.errors.ChildFailedError:

run_pretraining_multimae.py FAILED

Failures: <NO_OTHER_FAILURES>

Root Cause (first observed failure): [0]: time : 2022-07-09_16:11:15 host : Norwalk rank : 3 (local_rank: 3) exitcode : -6 (pid: 182201) error_file: <N/A> traceback : Signal 6 (SIGABRT) received by PID 182201

END______________

Thanks for the help

Hi, I think the following correction holds: Line 357 in https://github.com/EPFL-VILAB/MultiMAE/blob/main/run_finetuning_depth.py should be dataset_train = build_regression_dataset(args, data_path=args.train_data_path, transform=train_transform) instead of dataset_train = build_regression_dataset(args, data_path=args.data_path, transform=train_transform) or the argument train_data_path should be changed to data_path.

Apart from that, I am trying to recreate your results on NYUv2 for depth. but the dataset preparation instructions are not clear from the instructions in SETUP. As explained about the folder structure , where should the GT be when finetuning for depth and evaluating. Apart from that, mask_valid for fine-tuning? RuntimeError: Found 0 logs in subfolders of: /tmp-network/user/varora/multimae/multimae_data/train/mask_valid

Hi, Thank you for the interesting work and the extensive experiments. Your depth results are based on the DPT head in the paper. In the colab, you use the spatial adapter head for inference. I was wondering if your fine-tuning results with the spatial adapter head were better/worse than the DPT head? Was the intention to implement this spatial head more to test a pure transformer based head (compared to DPT's convolution based refineNet like approach?)?

Thank you.

HI, I find that:

in MultiMae/run_finetuning_semseg.py line735

seg_pred_argmax = seg_pred[:num_classes].argmax(dim=1) 

I think it should be

seg_pred_argmax = seg_pred[:,:num_classes,:,:].argmax(dim=1) 

Hello! I just want to setup this nice work quickly, and got a problem with inference. I used 'run_finetuning_depth.py' trying to see the code running well, also followed 'setup.md' to download NYUv2 dataset and structure folders. But got not found error 'mask_valid' folder. How can I make the 'mask_valid' folder for evaluate only?