|
@@ -133,8 +133,7 @@ Each caption is a list of words. During preprocessing, a dictionary is created
|
|
|
that assigns each word in the vocabulary to an integer-valued id. Each caption
|
|
that assigns each word in the vocabulary to an integer-valued id. Each caption
|
|
|
is encoded as a list of integer word ids in the `tf.SequenceExample` protos.
|
|
is encoded as a list of integer word ids in the `tf.SequenceExample` protos.
|
|
|
|
|
|
|
|
-We have provided a script to download and preprocess the [MSCOCO]
|
|
|
|
|
-(http://mscoco.org/) image captioning data set into this format. Downloading
|
|
|
|
|
|
|
+We have provided a script to download and preprocess the [MSCOCO](http://mscoco.org/) image captioning data set into this format. Downloading
|
|
|
and preprocessing the data may take several hours depending on your network and
|
|
and preprocessing the data may take several hours depending on your network and
|
|
|
computer speed. Please be patient.
|
|
computer speed. Please be patient.
|
|
|
|
|
|
|
@@ -262,8 +261,7 @@ tensorboard --logdir="${MODEL_DIR}"
|
|
|
### Fine Tune the Inception v3 Model
|
|
### Fine Tune the Inception v3 Model
|
|
|
|
|
|
|
|
Your model will already be able to generate reasonable captions after the first
|
|
Your model will already be able to generate reasonable captions after the first
|
|
|
-phase of training. Try it out! (See [Generating Captions]
|
|
|
|
|
-(#generating-captions)).
|
|
|
|
|
|
|
+phase of training. Try it out! (See [Generating Captions](#generating-captions)).
|
|
|
|
|
|
|
|
You can further improve the performance of the model by running a
|
|
You can further improve the performance of the model by running a
|
|
|
second training phase to jointly fine-tune the parameters of the *Inception v3*
|
|
second training phase to jointly fine-tune the parameters of the *Inception v3*
|
|
@@ -333,6 +331,4 @@ expected.
|
|
|
|
|
|
|
|
Here is the image:
|
|
Here is the image:
|
|
|
|
|
|
|
|
-<center>
|
|
|
|
|
|
|
|
|
|
-</center>
|
|
|
talkdirty