forked from videoflow/videoflow
-
Notifications
You must be signed in to change notification settings - Fork 0
/
TODO.txt
23 lines (22 loc) · 1.07 KB
/
TODO.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1. Documentation:
1.1 Creating a simple flow application
1.2 Starting and stopping the flow
1.2 Writing your own producers, processors and consumers
1.3 Checking the performance of the flow and identifying bottlenecks
1.4 Allocating more than multiple processes for a processor node.
1.4 Dealing with errors.
1.5 Reusing processors in multiple nodes in the graph.
1.6 Modifying the default behavior of the flow.
2. Code:
2.1 Add tensorflow processor and tracker processor
2.2 Add rest api processor
2.1 Add counting processor
2.2 Add ability to allocate more than one process for a processor
2.3 Write simple car counting flow for tracking and counting cars
2.4 Write code for outputing logs that analyze flow performance to identify bottlenecks
2.5 Add error handling and logging
2.6 Write ability to deploy computation nodes in other machines
2.7 Create tensorboard server that can read the logs of the flow and display them
in a screen.
2.8 Change the execution environment to use zeromq instead of queues and processes.
In that way, new nodes can join a flow after it has started running.