improved-diffusion
improved-diffusion copied to clipboard
openai
Release for Improved Denoising Diffusion Probabilistic Models
I downloaded CIFAR10 500K checkpoint from readme, sampled it for 10k images and my FID points to > 200. I tested few different implementations of FID calculations and am comparing...
For the npz file generated after sampling, how to calculate the FID score of the npz file and the cifar10 data set used ? Cifar is a picture in PNG...
Traceback (most recent call last): File "E:\learn_obj\fq_improved_diffusion\improved_diffusion\dist_util.py", line 73, in sync_params dist.broadcast(p, 0) File "D:\anaconda3\envs\pytorch2.0\lib\site-packages\torch\distributed\distributed_c10d.py", line 1451, in wrapper return func(*args, **kwargs) File "D:\anaconda3\envs\pytorch2.0\lib\site-packages\torch\distributed\distributed_c10d.py", line 1574, in broadcast work.wait() RuntimeError:...
Hello, As the paper explains, the authors experimented with different numbers of residual blocks per resolution level. It is my understanding that the number should be identical for the downsampling...
Hi, there! May I ask how long would the training normally takes for training on the CIFAR10 and ImageNet? I'm using 4x 16GB V100 GPUs I used the following settings...
I trained for a day using this command: ``` python scripts/image_train.py --data_dir datasets/cifar_train/ $MODEL_FLAGS $DIFFUSION_FLAGS $TRAIN_FLAGS ``` However, when I tried to sample: ``` python scripts/image_sample.py --model_path /tmp/openai-2022-08-14-03-40-40-000122/opt160000.pt $MODEL_FLAGS $DIFFUSION_FLAGS...
Sorry to bother you, but I am confused about the implementation of ddim_sample(), It seems that you re-derive the eps due to the possible clip operation to pred_x_start. I don't...
Hello, I am not sure of how to use the super resolution training ? What kind of dataset does it need ? Thank you so much
Thank you for your excellent job. I wonder which dataset is "Upsampling 256x256 model (280M parameters, trained for 500K iterations)" is trained on?
