score_sde_pytorch
score_sde_pytorch copied to clipboard
yang-song
PyTorch implementation for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)
https://github.com/yang-song/score_sde_pytorch/blob/cb1f359f4aadf0ff9a5e122fe8fffc9451fd6e44/sde_lib.py#L244 I think this line should be ``` -N / 2. * np.log(2 * np.pi * self.sigma_max ** 2) - torch.sum((z-x) ** 2, dim=(1, 2, 3)) / (2 * self.sigma_max...
--------------------------------------------------------------------------- ValueError Traceback (most recent call last) in 37 model=score_model, ema=ema) 38 ---> 39 state = restore_checkpoint(ckpt_filename, state, config.device) 40 ema.copy_to(score_model.parameters()) /workspace/utils.py in restore_checkpoint(ckpt_dir, state, device) 13 else: 14 loaded_state...
In the case where downsampling is true and dilation>1 there are two convolutions both of which are downsampling by the number of dilation. However the shortcut is only downsampled with...
Hello, In ./sampling.py, line 190, you have a class `ReverseDiffusionPredictor`: ``` @register_predictor(name='reverse_diffusion') class ReverseDiffusionPredictor(Predictor): def __init__(self, sde, score_fn, probability_flow=False): super().__init__(sde, score_fn, probability_flow) def update_fn(self, x, t): f, G = self.rsde.discretize(x,...
Hey guys, I found a stable environment that could run this code base. Here are my settings. System: WSL2 Ubuntu 22.04 GPU: NVIDIA A6000; NVIDIA-SMI 550.54.10; Driver Version: 551.61; CUDA...
Hi Song, In the paper you define $\alpha_i = \prod_{j=1}^i 1- B_j$ however in the actual implementation $\alpha_i = 1 - \beta_i$. Could you help to verify which one is...
(score_sde_pytorch) root@dqnjrsjllvraojle-wind-5b766c4d47-m5l9l:/data/coding/score_sde_pytorch# python3 main.py --config ./configs/ve/bedroom_ncsnpp_continuous.py --mode train --workdir score_sde/ve 2025-01-17 18:24:03.674485: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off...
Hello, when I training the code in my single RTX-4090, it said that OOM. Even I set the batch size to 2, it still has this problem. Anyone know how...
As some of the other people commented, the requirements file doesn't work and has conflicts with its dependencies, regardless of Python version. Also the demo "Score_SDE_demo_PyTorch.ipynb" can't be run because...
