-
Hi, I am running the test files for Neural Tangents and a lot of cases were skipped. For example, 96 / 127 test cases were skipped in Neural Tangents stax test cases. I looked at the implementation of…
-
Hi, I reinstalled some packages and I reran the tests of Neural Tangents (latest version). But I am getting an interesting error and have not found a solution. Previously I ran Neural Tangents tests a…
-
As titled, seems to be similar to issue #15 , but have not figured out the exact reason for this one.
```
stax.py", line 3405, in
masks = [m if m is None else np.broadcast_to(
File "/Libr…
-
Hi, thanks for your effort on developing this library. However, when I conduct some experiments, the example code of WideResNet outputs NaN on MNIST dataset but seems normal on cifar10 dataset.
My t…
ybj14 updated
4 years ago
-
## In the last part of [5-minute-intro](https://github.com/google/neural-tangents#5-minute-intro)
I guess the line assigning `y_test_ntk` should pass parameter `get='ntk'` instead of `get='nngp'`
…
-
Hi,
It seems to me that when evaluated at input data, the analytic NTK dimensions are not consistent with empirical NTK dimensions. Concretely, consider a small MLP as the follows:
```python
f…
-
When installing using `pip3` on my PC, the bug in line 421 of `neural_tangents/utils/batch.py` still exists (i.e., `np.onp`), although it seems to have already been revised in commit [https://github.c…
-
Hi, could anyone explain to me about the `` object defined in [https://github.com/google/jax/blob/master/jax/interpreters/partial_eval.py](https://github.com/google/jax/blob/master/jax/interpreters/pa…
-
Hello,
I am reading [your paper](https://arxiv.org/abs/2006.14548) but I am not aware of the training/resources cost of this kind of kernel. Is it faster or slower to train than regular networks? I…
-
Hi, sorry for bothering. In the `test_composition_conv_avg_pool` test cases, some outer products on the covariance matrices are performed while doing Kernel transformation. In the outer product functi…