Hao Phung

I met this error but we can solve it by uncommenting some codes in `base_trainer.py` from line 75-76: ``` if len(self.opt.gpus) > 1: model_with_loss = self.model_with_loss.module model_with_loss.eval() torch.cuda.empty_cache() ```

Thank you for your response! I finally understand your points after re-checking section 4 in your paper. Btw, have you experimented with a sampled latent from VAE for the generative...

Have you downloaded pretrained `stable-diffusion-2-1-base` and put it in `stable-diffusion` directory? If not, please download it first following our [Environment setup](https://github.com/VinAIResearch/Anti-DreamBooth/blob/main/README.md#environment-setup) and then run our script. Thanks!

The use of pretrained autoencoder is to enhance training efficiency and performance of the model. Hence, `first_stage_model` here is only performed in inference mode, without further training. In case, you...

The training is relatively stable and consistent for each single experiment. It is not as varied as your provided results. May you provide us your training hyper-params and the detail...

I think you misunderstand here. To compute FID, following standard practice, you should generate 50_000 images for statistical significance.

50_000 is the number of output images that you use your trained model at epoch 450 `model_450.pth` to generate, not the number of generated images during training. To test, you...

It is weird because we experience no problem with inference code. I have tried to clone new repo and reinstall all of the required libs and checkpoint models. ``` root@6b3b847909fd:~/LFM#...

thank you for asking! I hope it is not too late, here is the link for all ADM-based exps https://drive.google.com/drive/folders/1looUTAvOdBMQRvlsavK0rxDhdzdhbToK?usp=sharing

Hi folk, We juse jpg format and generate 50K images for eval. For down sampling on FFHQ, we use transforms.Resize(args.image_size) (see [here](https://github.com/VinAIResearch/LFM/blob/50d263b0a4471384c0ab073c8f904b73c024c6c0/datasets_prep/__init__.py#L104)).