Here a colab that demostrates the bug
https://colab.research.google.com/drive/1OWSEoV7Wx-EBA_2IprNZoASNvXIky3iC?usp=sharing
++++the following code gives"received an invalid combination of arguments - got (Tensor, Tensor), but expected one of:" 
import torch
torch.nn.Linear(
torch.as_strided(torch.ones((4,6*5*5*70*10)),(4,6*5*5,10),(4,2,-1))
,torch.ones(10,6*5*6))
#(4,2,3),(2,3,1))
++++The following code gives "TypeError: ones(): argument 'size' (position 1) must be tuple of ints, not str import torch"
import torch
#torch.nn.Linear(
torch.as_strided(torch.ones((4,6*5*5*70*10)),(4,6*5*5,10),(4,2,-1))
,torch.ones(10,6*5*6)#)
#(4,2,3),(2,3,1))
++++

This doesn't appear to be a Python bug, but rather an issue with how you're using PyTorch. I suggest you ask for help on a PyTorch support forum, or maybe try StackOverflow.

This appears to be a bug with the google colab site.

For whatever reason, if you try to evaluate a statement that is a line with  a leading comma (afaik, never valid python), then colab does something wierd by wrapping the arguments in quotes, and evaluating them ...

Here, `>>>` denotes running python in a colab cell:

>>> type(1)
int
>>> ,type(1)
str
>>> ,type(1,2,3)
str
>>> ,type(1, 2, 3)
---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-38-a1b277d7db3e> in <module>()
----> 1 type("(1,", "2,", "3)")

TypeError: type.__new__() argument 2 must be tuple, not str

---

This is not reproducible in normal python, so seems extremely likely to be some (planned or unplanned) feature of google colab that's causing confusion here.

I'm going to close this as not a python bug. If someone gets some evidence to the contrary, we can re-open it.