hi, i met with a problem when trying to convert encoder to onnx for deploy. errors pop as follow: torch.onnx.symbolic_registry.UnsupportedOperatorError: Exporting the operator prim::layout to ONNX opset version 16 is not supported. Please feel free to request support or submit a pull request on PyTorch GitHub.

after a careful debug, i cannot locate the exact usage of torch::Layout.

thanks!

I follow the step-by-step instructions of your article to repeatedly install and uninstall in the Windows 10 environment: Anaconda3、cuda10.1、python-3.8.5、 pytorch1.7.1、mmcv-full1.4.2，

However, when running the command: python demo/quick_ Try.py demo/src.png -- style toy is always stuck in an error:

Fail to import MultiScaleDeformableAttention from mmcv.ops.multi_ scale_ deform_ attn, You should install mmcv-full if you need this module.

In addition, I have installed mmcv according to your article guidance, also according to Baidu method, and downloaded whl (Windows version) directly on the official website and then installed it with instructions, However, the problem remains. How can I solve this problem? thank you.

Here are the instructions I often use:

conda create -n facestylor python=3.8.5 -y conda activate facestylor conda install pytorch=1.7.1 cudatoolkit=10.1 torchvision -c pytorch

pip install mmcv-full==1.4.2 -f https://download.openmmlab.com/mmcv/dist/cu{Cuda_verison}/torch{Pytorch_Version}/index.html

pip install mmcv-full==1.4.2 -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7.1/index.html

git clone https://github.com/open-mmlab/mmgeneration.git cd mmgeneration

pip install -r requirements.txt pip install -v -e .
cd ..

运行：H:\Program Files\MMGEN-FaceStylor python demo/quick_try.py demo/src.png --style toonify

遇到问题： 1、No module named 'cv2' pip install opencv-contrib-python

2、You should install mmcv-full if you need this module pip install -U openmim mim install mmcv-full

File "/data/miniconda3/envs/facestylor/lib/python3.7/site-packages/mmcls/init.py", line 57, in f'MMCV=={mmcv.version} is used but incompatible. '
AssertionError: MMCV==1.3.15 is used but incompatible. Please install mmcv>=1.4.2, <=1.9.0.

Thank you for making this repository.

It might be hard to run custom trainning due to some issues in the repo. Here are tips to make it work:

Env: RTX 3080, Ubuntu 18.04, Python 3.8, Pytorch 1.7.0, CUDA 11, CUDAToolkit 11.0.221, mmvc(torch1.7.0 and cu110), mmgeneration(f6551e1)

1. Issue related to mmgeneration SiLU is registered

Comment line 35(the annotation to register SiLU) in mmgeneration/mmgen/models/architectures/ddpm/modules.py, then install mmgeneration.

2. Build error when compiling custom operators nvcc fatal : Unsupported gpu architecture ‘compute_86‘ on new graphics card series.

Use export TORCH_CUDA_ARCH_LIST="8.0" before running.

3. Pretrained weights is missing.

Download pretrained weights in configs into work_dirs/pre-trained/ before running.

4. Running trainer directly with Python.

Add agilegan folder in tools/train.py,

import sys
sys.path.append(PATH_TO_MMGEN-FaceStylor)
import agilegan  # isort:skip  # noqa

then use python tools/train.py PATH_TO_YOUR_CONFIGS --work-dir PATH_TO_YOUR_DIR --gpus GPU_NUMS.

5. object has no attribute 'module' in transfer.py and log_buffer boardcasting error

Fix agilegan/transfer.py with, ~line 140 and 210

# obtain some params
        # fix
        # g_log_size = self.generator.module.log_size
        g_log_size = self.generator.log_size
        # fix
        # if hasattr(self.generator.module, 'num_layers'):
        if hasattr(self.generator, 'num_layers'):
            # fix
            # g_num_layers = self.generator.module.num_layers
            g_num_layers = self.generator.num_layers
        else:
            # fix
            # g_num_layers = self.generator.module.num_injected_noises
            g_num_layers = self.generator.num_injected_noises
        # fix
        # d_log_size = self.discriminator.module.log_size
        d_log_size = self.discriminator.log_size

~line 460

# update ada p
        # fix
        # if hasattr(self.discriminator.module,
        #            'with_ada') and self.discriminator.module.with_ada:
        if hasattr(self.discriminator,
                   'with_ada') and self.discriminator.with_ada:
            # self.discriminator.module.ada_aug.log_buffer[0] += 1
            self.discriminator.ada_aug.log_buffer[0] += 1
            # self.discriminator.module.ada_aug.log_buffer[
            self.discriminator.ada_aug.log_buffer[
            # fix
            #    1] += disc_pred_real.sign()
               1] += disc_pred_real.sign().sum()
            # self.discriminator.module.ada_aug.update(iteration=curr_iter,
            self.discriminator.ada_aug.update(iteration=curr_iter,
                                                     num_batches=batch_size)
            log_vars_disc['ada_prob'] = (
                # self.discriminator.module.ada_aug.aug_pipeline.p.data)
                self.discriminator.ada_aug.aug_pipeline.p.data)

6. Issue related to logger, gpu data should use cpu() before numpy().

Disable logger in your config, you can get info in work directory.

# log_config = dict(interval=100, hooks=[dict(type='TextLoggerHook')])
log_config = dict(interval=100, hooks=[])

Hopes this helps.

Hi, I try to train a new encoder but it has some errors: AttributeError: 'SwapStyleGANv2Generator' object has no attribute 'module'. How to solve like this? @plyfager @TommyZihao

We attempt to train a new encoder, but it's shown this errors as follow.

Traceback (most recent call last): File "tools/train_encoder.py", line 161, in main() File "tools/train_encoder.py", line 138, in main test_cfg=cfg.test_cfg) File "/root/picasso/haoyuan_chen/chy/mmgeneration/mmgen/models/builder.py", line 32, in build_model return build(cfg, MODELS, dict(train_cfg=train_cfg, test_cfg=test_cfg)) File "/root/picasso/haoyuan_chen/chy/mmgeneration/mmgen/models/builder.py", line 27, in build return build_from_cfg(cfg, registry, default_args) File "/root/picasso/haoyuan_chen/anaconda3/envs/facestylor/lib/python3.7/site-packages/mmcv/utils/registry.py", line 45, in build_from_cfg f'{obj_type} is not in the {registry.name} registry') KeyError: 'PSPEncoderDecoder is not in the model registry'

And we refer to the methods on the forum, such as reconfiguring the requirements.txt and using @models register_ module(name='PSPEncoderDecoder', force=True), but they are useless.

We now suspect that it is caused by the mismatch between the version of mmcv and the version of mmgen. Our version of mmcv is 1.4.8 and the version of mmgen is 0.6.0.