Comments for Hack Till Dawn

Comment on Inception modules: explained and implemented by How to build to a Fashion Classifier in 5 easy steps using Deep Learning. - AI+ NEWS

Fri, 15 Feb 2019 18:49:59 +0000

[…] https://hacktilldawn.com/2016/09/25/inception-modules-explained-and-implemented/ […]

Comment on Inception modules: explained and implemented by Superhero Toolkit for Data Science - Superhero Adventures in Data Science!

Superhero Toolkit for Data Science - Superhero Adventures in Data Science! — Thu, 31 Jan 2019 21:10:40 +0000

[…] that are also playing important roles in winning competitions and solving problems, too. ResNet, Inception Modules (models within models) and Highway Networks are allowing the Superheroes to keep their gradients […]

LikeLike

Comment on Inception modules: explained and implemented by abou

abou — Fri, 30 Nov 2018 16:48:20 +0000

Hi, thank you for your tutorial. it is well explained. When I run you notebook, i get the following error. Can you please help me? I have already search on the net but I can’t find a solution. Thanks

ValueErrorTraceback (most recent call last)
in ()
138
139 loss = tf.reduce_mean(
–> 140 tf.nn.softmax_cross_entropy_with_logits(model(X),y_))
141 opt = tf.train.AdamOptimizer(1e-4).minimize(loss)
142

in model(x, train)
110
111 #concatenate all the feature maps and hit them with a relu
–> 112 inception1 = tf.nn.relu(tf.concat(3,[conv1_1x1_1,conv1_3x3,conv1_5x5,conv1_1x1_4]))
113
114

/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/array_ops.pyc in concat(values, axis, name)
1120 axis, name=”concat_dim”,
1121 dtype=dtypes.int32).get_shape().assert_is_compatible_with(
-> 1122 tensor_shape.scalar())
1123 return identity(values[0], name=scope)
1124 return gen_array_ops.concat_v2(values=values, axis=axis, name=name)

/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/tensor_shape.pyc in assert_is_compatible_with(self, other)
846 “””
847 if not self.is_compatible_with(other):
–> 848 raise ValueError(“Shapes %s and %s are incompatible” % (self, other))
849
850 def most_specific_compatible_shape(self, other):

ValueError: Shapes (4, 50, 28, 28, 32) and () are incompatible

LikeLike

Comment on Inception modules: explained and implemented by What's my name?

What's my name? — Wed, 17 Oct 2018 23:59:12 +0000

Wish I had found this earlier. Such awesomeness. Thanks man.

LikeLike

Comment on Inception modules: explained and implemented by sanderali

sanderali — Tue, 09 Oct 2018 07:57:34 +0000

In reply to Dinesh Vadhia.

Dinesh Vadhia, have u tried the ReLus to get a sparse output?

LikeLike

Comment on Inception modules: explained and implemented by Major Resources used for CNN Part – II – Deep Learning

Major Resources used for CNN Part – II – Deep Learning — Wed, 05 Sep 2018 05:07:11 +0000

[…] 6. https://hacktilldawn.com/2016/09/25/inception-modules-explained-and-implemented/ […]

LikeLike

Comment on Vanilla Recurrent Neural Networks by Ugenteraan

Ugenteraan — Sat, 21 Jul 2018 08:47:53 +0000

Thank you very much for this.

LikeLike

Comment on Using your remote Jupyter Notebook through ssh by Alex Dai

Alex Dai — Mon, 11 Jun 2018 18:03:40 +0000

Any reason why calls to %qtconsole don’t work when I use this method? Thanks

LikeLike

Comment on Inception modules: explained and implemented by priya

priya — Sun, 20 May 2018 11:05:54 +0000

if we apply 1×1 convolution on 192@28*28, how does that outputs a 16 28×28 feature maps and on 16*28*28, if we do the 5×5 convolutions on those feature maps how does that outputs 32 28×28 feature maps. Can you please explain this ?

LikeLike

Comment on Inception modules: explained and implemented by Ankush Manocha

Ankush Manocha — Tue, 03 Apr 2018 06:07:45 +0000

Awesome Job Sir, Really Appreciate.

LikeLike