Resource retrieval

Get the pre-trained net:

In[1]:=

Out[1]=

Basic usage

Record an audio sample and transcribe it:

In[2]:=

In[3]:=

In[4]:=

Out[4]=

Join the characters:

In[5]:=

Out[5]=

Display the confidence of the five most likely hypotheses:

In[6]:=

Out[6]=

Performance evaluation

Models trained with CTC loss have difficulties in producing correct spelling:

In[7]:=

Out[7]=

The standard solution to this problem is to incorporate a language model into the beam search decoding procedure, as done in the original Deep Speech 2 implementation.

Noisy environments can affect the performance too:

In[8]:=

Out[8]=

Visualization

Write a function to visualize the token probabilities for each time step:

In[9]:=

Inspect the character probabilities of the sample. The <blank> token is the standard blank token used by nets trained with CTC loss:

In[10]:=

Out[10]=

Net information

Inspect the sizes of all arrays in the net:

In[11]:=

Out[11]=

Obtain the total number of parameters:

In[12]:=

Out[12]=

Obtain the layer type counts:

In[13]:=

Out[13]=

Display the summary graphic:

In[14]:=

Out[14]=

Export to MXNet

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

In[15]:=

Out[15]=

Export also creates a net.params file containing parameters:

In[16]:=

Out[16]=

Get the size of the parameter file:

In[17]:=

Out[17]=

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

In[18]:=

Out[18]=