tangent
tangent copied to clipboard
Error in tangent tutorial notebook: "AttributeError: module 'gast' has no attribute 'Num'"
Hi i opened your tutorial notebook and run the cells in notebook. after installing tangent at cell with following code:
import tangent
df = tangent.grad(f)
the following error arises:
AttributeError Traceback (most recent call last)
3 frames
/usr/local/lib/python3.6/dist-packages/tangent/grammar.py in
AttributeError: module 'gast' has no attribute 'Num'
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
This yield the issue:
ERROR: Could not find a version that satisfies the requirement gast==0.2.4 (from versions: 0.1, 0.1.1, 0.1.2, 0.1.3, 0.1.4, 0.1.5, 0.2.0, 0.2.1, 0.2.1.post0, 0.2.1.post1, 0.2.2, 0.3.0, 0.3.1, 0.3.2, 0.3.3, 0.4.0, 0.5.0, 0.5.1, 0.5.2, 0.5.3)
ERROR: No matching distribution found for gast==0.2.4
Looks like there's no gast 0.2.4.
Using 0.3.0 yields
[/usr/local/lib/python3.7/dist-packages/tangent/grammar.py](https://localhost:8080/#) in <module>()
16 import gast
17
---> 18 LITERALS = (gast.Num, gast.Str, gast.Bytes, gast.Ellipsis, gast.NameConstant)
19
20 CONTROL_FLOW = (gast.For, gast.AsyncFor, gast.While, gast.If, gast.Try)
AttributeError: module 'gast' has no attribute 'Num'
while using 0.2.2 yields
[/usr/local/lib/python3.7/dist-packages/tangent/tf_extensions.py](https://localhost:8080/#) in <module>()
58
59 non_differentiable.register_non_differentiable_functions(
---> 60 tf.shape, tf.to_float, tf.equal, tf.constant,
61 tf.zeros, tf.ones, tf.zeros_like, tf.ones_like,
62 size, shape_as_list, dtype)
AttributeError: module 'tensorflow' has no attribute 'to_float'
All these examples are from the Colab notebook. It's such a shame Tangent isn't functional anymore, it would have been pretty useful.
Odd that that version is gone. Does 0.2.2 work instead?
No, both 0.3.0 and 0.2.2 do not work. My previous comment contains errors for both versions!
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.
I took a stab at upgrading to gast 0.3.0 in smbarr/tangent-upgrade-gast
Most of the tests seem to be passing now, but I'm having a lot of trouble getting some of the tensorflow derivatives to work.
I would really like to get all the tests passing, so if anyone has an idea of how to do that I'm happy to help further. This is a really interesting and unique project which I'd hate to see die.