Wolfram Computation Meets Knowledge

Inception V1 Trained on Places365 Data

Identify the scene type of an image

Originally released in 2016 by the MIT Computer Science and Artificial Intelligence Laboratory as a pre-trained model for the launch of the Places365 dataset (a subset of the Places2 dataset). The model is based on the Inception V1 architecture and achieves 84.01% top-five (53.59% top-one) accuracy on the Places365 dataset.

Number of layers: 147 | Parameter count: 6,347,677 | Trained size: 26 MB

Training Set Information

Examples

Resource retrieval

Retrieve the resource object:

In[1]:=
ResourceObject["Inception V1 Trained on Places365 Data"]
Out[1]=

Get the pre-trained net:

In[2]:=
NetModel["Inception V1 Trained on Places365 Data"]
Out[2]=

Basic usage

Find the type of scene of an image:

In[3]:=
Out[3]=

Obtain the probabilities of the most likely scenes:

In[4]:=
Out[4]=

Export to MXNet

Export the net into a format that can be opened in MXNet:

In[5]:=
jsonPath = 
 Export[FileNameJoin[{$TemporaryDirectory, "net.json"}], 
  NetModel["Inception V1 Trained on Places365 Data"], "MXNet"]
Out[5]=

Export also creates a net.params file containing parameters:

In[6]:=
paramPath = FileNameJoin[{DirectoryName[jsonPath], "net.params"}]
Out[6]=

Get the size of the parameter file:

In[7]:=
FileByteCount[paramPath]
Out[7]=

The size is similar to the byte count of the resource object:

In[8]:=
ResourceObject["Inception V1 Trained on Places365 Data"]["ByteCount"]
Out[8]=

Represent the MXNet net as a graph:

In[9]:=
Import[jsonPath, {"MXNet", "NodeGraphPlot"}]
Out[9]=

Requirements

Wolfram Language 11.1 (March 2017) or above

Reference