-
The following script fails with Keras 1.0.4, but worked with 1.0.3:
```
from keras.layers import Dense, Activation
from keras.models import Sequential
model = Sequential([
Dense(32, input_dim=2),…
-
GpuManagedAllocator was early supported at tensorflow/core/common_runtime/gpu/gpu_managed_allocator.cc.
But it seems there is no choice for users to use it according to the source code?
The GpuManag…
-
Hi,
When trying to apply the GroupNormalization Layer to fully convolutionnal networks with a dynamic input shape (as there are not mathematical constraints for a fixed input shape and keras usuall…
-
See https://github.com/pytorch/pytorch/issues/1601 for previous discussion on [layer normalization](https://arxiv.org/abs/1607.06450).
-
It useful to allow TensorFlow functions to be parametrized by e.g. the dimensions of the weight matrices and sizes of convolutions. It is natural to treat these parameters statically - they do not ch…
-
The current implementation of `tf.contrib.layers.batch_norm` is quite slow with the default parameters. Recent builds of TF support `fused=True` which forces the use of the faster `nn.fused_batch_norm…
-
I'm running tensorflow on a machine with 64 CPUs and no GPUs. I noticed that although tf detects the number of CPU's correctly (
I tensorflow/core/common_runtime/local_device.cc:40] Local device intra…
-
to foster community involvement - some richer sample code beyond MNIST should be tackled.
Generative Adversarial Networks is a hot topic amongst ML - and some sample code using swift should help enco…
-
Based on some recent benchmarks (on GTX 1080 Ti), using 8-bit quantized integers during inference could give us significant performance improvement. No experiments has been performed yet, but other so…
-
The relu example is following:
https://github.com/facebookresearch/TensorComprehensions/blob/9cc088fe034142199fd63717a2d59b15394b38c4/test_python/layers/test_relu.py#L25-L33
In this example, the i…