-
hi, thanks for the great work here. I tried to run the pretrained model but I get this error:
Using GPU 0
['Tesla K80']
{'cuda': 0, 'comment': 0, 'batch_size': 8, 'train_data_dir': '/content/drive/…
-
Thank you for sharing the code. Could you please point me to where the UV Total Variation Loss mentioned in the paper is implemented in the code?
-
# 🐛 Bug
I am using the sparse variational GPyTorch framework to perform 7500 tasks. I have 4800 data points, and I am using batch sizes (so both the inout and output matrices have dimension (4800…
-
To add a loss and metrics to a model, I can add them to `model.compile(loss=..., metrics=...)`, provided that they have the signature `fn(y_true, y_pred)`, see the [docs](https://keras.io/api/models/m…
-
Thanks for this great work, however, I still have some questions about training strategy and loss term.
The training code of appearance generative network in `train.sh` is as below:
https://github…
-
In the documentation for variational Gaussian process applied to minibatches (https://github.com/tensorflow/probability/blob/v0.12.1/tensorflow_probability/python/distributions/variational_gaussian_pr…
-
`## Basic configuration
style_image: img/snow.jpg # targeted style image
naming: "snow" # the name of this model. Determine the path to save checkpoint and events file.
model_path: models # root p…
-
So I've been working on a very long meshed highway network with a vgg-18 perceptual loss to boot. However, adding just a few more commands for a RGB to HSV conversion in the loss causes the whole thi…
-
You normalize the data in (-1,1) and the parameter of tf. log () is greater than 0, so loss is nan.
-
MNIST Results:
We used a standard variational autoencoder. We trained with inlier digits 1 and 3. Each model was trained with batch size of 128, where each image is an unnormalized grayscale image of…