enumerate()
Objects that support iteration
The enumeration object generates pairs containing a count (from start, which defaults to 0) and a value generated by the iterable parameter.
Enumerate is useful for getting a list of indices:
unique()
Get the unique elements in the array.
map()
df[5] = df[4].map(mapping)
Use the value in df[4] as the key, and get a list of values stored in df[5].
torch.rand()
Returns the initialized tenosr that obeys the uniform distribution, and its shape is its parameter size
U = torch.rand([m,rank], requires_grad = True, dtype=torch.float)
torch.nn.functional.cross_entropy()
torch.nn.functional.cross_entropy (self-calculated value, target value)
Calculate cross entropy
H(p,q)=−i∑P(i)logQ(i)