improved-diffusion
improved-diffusion copied to clipboard
openai
Release for Improved Denoising Diffusion Probabilistic Models
OS: Ubuntu 22.04 Graphic: RTX 3090 Python 3.10 mpi4py: 3.5.1 train_bash.sh ```shell #!/bin/bash MODEL_FLAGS="--image_size 32 --num_channels 128 --num_res_blocks 3 --learn_sigma True --dropout 0.3 --class_cond True " DIFFUSION_FLAGS="--diffusion_steps 4000 --noise_schedule cosine"...
Thanks for your great work. I try to train a diffusion model on cifar-10. I used the following parameters: ``` MODEL_FLAGS="--image_size 32 --num_channels 128 --num_res_blocks 3 --learn_sigma True --dropout 0.3"...
when set fp16=True, there is an error about `_unflatten_dense_tensors`: ``` TypeError: unflatten_dense_tensors(): argument 'tensors' (position 2) must be tuple of Tensors, not generator ``` Is this due to the version...
when i try to run image_sample.py, i meet this question.Have anyone meet the same question? raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for UNetModel: Missing...
I want to use this github repository to train a DDPM model based on CelebA dataset consisting of 202,599 align&cropped human face images (each image is 218(height)x178(width) pixels) for human...
What does "base_samples" mean when I sample with a trained model? When I try to enter the image that I want to do super resolution, I get an error:OSError: Failed...
This is a great code for the diffusion model. I trained for over 10,000 steps with Colab. Will the training stop automatically? And I can't find the .pt file. And...
python scripts/image_sample.py --model_path /hd8t/user_code/lxr/improved-diffusion/ckpt_save/ema_0.9999_025000.pt $DIFFUSION_FLAGS $MODEL_FLAGS RuntimeError: Error(s) in loading state_dict for UNetModel: Missing key(s) in state_dict: "input_blocks.7.0.skip_connection.weight", "input_blocks.7.0.skip_connection.bias", "input_blocks.10.0.skip_connection.weight", "input_blocks.10.0.skip_connection.bias", "input_blocks.10.1.norm.weight", "input_blocks.10.1.norm.bias", "input_blocks.10.1.qkv.weight", "input_blocks.10.1.qkv.bias", "input_blocks.10.1.proj_out.weight", "input_blocks.10.1.proj_out.bias", "input_blocks.11.1.norm.weight", "input_blocks.11.1.norm.bias", "input_blocks.11.1.qkv.weight",...
Hello, I am having an issue in using mpiexec to distribute the training. It seems that I can run training on a single GPU using the following parameters: `MODEL_FLAGS="--image_size 256...
When the image_dataset loads pictures, it will automatically convert the pictrue to rgb. However, i just want to keep the original type
