-
Notifications
You must be signed in to change notification settings - Fork 0
/
python_debug_test.py
37 lines (30 loc) · 1.03 KB
/
python_debug_test.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
from viz.debug import set_trace
import viz.graphtracker as gt
from sandbox.vgg import vgg16
from sandbox.stack_lstm import StackLSTM, PseudoLogLSTM
from torch.autograd import Variable
import torch
myInt = 86
myFloat = 3.14159
myBool = True
myString = "The quick brown fox jumps over the lazy dog"
myNone = None
myList = [1, 2.3, False, "hello", None, [10, 11, ["This", "is", "the", "end"]]]
myDict = {"key1": "value1"}
myTensor1 = (torch.rand(15,20) - 0.5) * 10
myTensor2 = torch.rand(6,7) - 0.5
myTensor3 = (torch.randn(100, 100))
def myFn(arg1):
return arg1 + 5
myVGGInput = gt.track_data(Variable(torch.ones(1, 3, 32, 32)), {})
myVGG = vgg16()
myVGGOutput = myVGG(myVGGInput)
myRNNBatchSize = 5
myRNNDims = [10, 10, 10, 10]
myRNNInput = [gt.track_data(Variable(torch.ones(5, myRNNDims[0])), {'self': None}) for _ in range(4)]
myRNN = StackLSTM(myRNNBatchSize, myRNNDims)
myRNNOutput = myRNN(myRNNInput)
myPseudoLogLSTM = PseudoLogLSTM(myRNNBatchSize, myRNNDims)
myPseudoLogLSTMOutput = myPseudoLogLSTM(myRNNInput)
set_trace()
print('Goodbye!')