Dan Moldovan
Dan Moldovan
Do you have the original code and arguments you used for gradients? A possible cause might be that `tangent.grad` assumes scalar output - can you try to see if using...
I see - the error originates from `np.dot`, the matrix sizes don't align properly. For correct matrix multiplication, the call should be `xxx(np.random.rand(1, 3), np.random.rand(1, 3))`, as you mentioned. For...
Yes, it seems that we have a bug in the handling of slice operators. This might be insufficient for your immediate needs, but it should run: ``` def test(x, y):...
Opened #62 for the slice issue.
Enumerate is not well supported right now (though we should have raised an error about that, so it's a bug either way). Using a regular walk `for i in range(len(x))`...
Yep, looks good now. The tests are unfortunately broken but #98 should fix that.
This is due to incompatible dependencies. We haven't had the chance to upgrade the code. Could you try installing an older version of gast: `pip install gast==0.2.4`
Odd that that version is gone. Does `0.2.2` work instead?
That means that the code would need to be upgraded to the newest version. Not terribly difficult, it's mostly mechanical, if anyone has the bandwitdth to do it.
Sure! Feel free to assign the issue to yourself and send me a PR. The fix should be straightforward, by replacing `tf.to_float` with tf.cast`.