Project3_1_ResNet_Transfer_Learning

*Goal
Training accuracy 80% for each Label

*Due
9/18~9/20

*Q&A
1.download pretrained net

Q

AttributeError: module 'urllib' has no attribute 'urlretrieve'

A 

From Python3 urllib module has been split into several modules. 

import urllib.request
data = urllib.request.urlretrieve("http://...") 

2.DATA Set Link
http://people.idsia.ch/~guzzi/DataSet.html

3.Logits meaning
 A.

loss_function = tf.nn.softmax_cross_entropy_with_logits(
     logits = last_layer,
     labels = target_output
)

Logit is a function that maps probabilities ([0, 1]) to R ((-inf, inf))

Probability of 0.5 corresponds to a logit of 0. Negative logit correspond to probabilities less than 0.5, positive to > 0.5.

(https://stackoverflow.com/questions/41455101/what-is-the-meaning-of-the-word-logits-in-tensorflow)

4.ResourceExhaustedError

5.list index error

>>> a = [1,2,3]
>>> a[4]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
IndexError: list index out of range
 

Reference 

https://github.com/tensorflow/models/tree/master/slim

https://github.com/sjchoi86/advanced-tensorflow/blob/master/cam/pretrain_usage.ipynb


https://kakalabblog.wordpress.com/2017/04/03/softmax-%ec%a0%95%eb%a6%ac/