Neural Network (Keras)
Keras, the high-level interface to the TensorFlow machine learning library, uses Graphviz to visualize how the neural networks connect. This is particularly useful for non-linear neural networks, with merges and forks in the directed graph.
This is a simple neural network (from Keras Functional
API) for ranking customer issue
tickets by priority and routing to which department can handle the ticket.
Generated using Keras' model_to_dot
function.
This model has three inputs:
- issue title text
- issue body test
- issue tags
and two outputs:
- predicted priority
- predicted department
Each node is labelled with the shape (length, width) of its input and output
matrices. None is shown where the shape is undecided yet, where the shape
depends on the final data you train this model against.
[Input .gv File] [SVG] [Raster Image]
digraph G {
fontname="Helvetica,Arial,sans-serif"
node [fontname="Helvetica,Arial,sans-serif"]
edge [fontname="Helvetica,Arial,sans-serif"]
concentrate=True;
rankdir=TB;
node [shape=record];
140087530674552 [label="title: InputLayer\n|{input:|output:}|{{[(?, ?)]}|{[(?, ?)]}}"];
140087537895856 [label="body: InputLayer\n|{input:|output:}|{{[(?, ?)]}|{[(?, ?)]}}"];
140087531105640 [label="embedding_2: Embedding\n|{input:|output:}|{{(?, ?)}|{(?, ?, 64)}}"];
140087530711024 [label="embedding_3: Embedding\n|{input:|output:}|{{(?, ?)}|{(?, ?, 64)}}"];
140087537980360 [label="lstm_2: LSTM\n|{input:|output:}|{{(?, ?, 64)}|{(?, 128)}}"];
140087531256464 [label="lstm_3: LSTM\n|{input:|output:}|{{(?, ?, 64)}|{(?, 32)}}"];
140087531106200 [label="tags: InputLayer\n|{input:|output:}|{{[(?, 12)]}|{[(?, 12)]}}"];
140087530348048 [label="concatenate_1: Concatenate\n|{input:|output:}|{{[(?, 128), (?, 32), (?, 12)]}|{(?, 172)}}"];
140087530347992 [label="priority: Dense\n|{input:|output:}|{{(?, 172)}|{(?, 1)}}"];
140087530711304 [label="department: Dense\n|{input:|output:}|{{(?, 172)}|{(?, 4)}}"];
140087530674552 -> 140087531105640;
140087537895856 -> 140087530711024;
140087531105640 -> 140087537980360;
140087530711024 -> 140087531256464;
140087537980360 -> 140087530348048;
140087531256464 -> 140087530348048;
140087531106200 -> 140087530348048;
140087530348048 -> 140087530347992;
140087530348048 -> 140087530711304;
}