-
I believe this formula is off by a sign
The correct formula should be the following since you took negation when defining the loss function (see also [the blog post from which you borrowed Fig.…
-
Hi Dan,
Sorry to bother again. I've successfully trained a network using jukebox perceptual loss that scores high on both MS-SSIM and MSE metrics (see below). The same occurs when training for long…
-
Song Y, Ermon S. [Generative modeling by estimating gradients of the data distribution](https://proceedings.neurips.cc/paper/2019/file/3001ef257407d5a371a96dcd947c7d93-Paper.pdf).
Yang Song. [Generat…
-
### Week 1 - Get to know the community
- [X] Join the communication channels
- [X] Open a GitHub issue (this one!)
- [X] Install the Ersilia Model Hub and test the simplest model
- [X] Write a m…
-
This project can be applied to VAE?
-
##### System information
- OpenVINO => 2022.2.0
- Operating System / Platform => Linux 64 Bit
- Compiler => g++
- Problem classification: Model Conversion
- Framework: PyTorch
- OS: Linux f93ecc…
-
Hi,
I'm new to both `JAX` and `diffrax`.
I'm familiar with latent SDE in Pytorch version `torchsde`. I find difficult to implement in `diffrax`.
Latent SDE has an augmented term which com…
-
Hello.
I tried VAE-GAN training with my data (256×256pix).
However , after typing `./scripts/train.sh ./img_data/data1` displayed following errors.
![Screenshot from 2021-12-16 20-16-50](https://…
-
Song Y, Ermon S. [Generative modeling by estimating gradients of the data distribution](https://arxiv.org/pdf/1907.05600.pdf).
-
GanttStart: 2022-01-31
GanttDue: 2022-02-01
After the initial research of the possible solutions is done a decision on the NN architecture shall be done.