I want to use tensorflow to get the full computation graph (including forward, backward and parameter update). I tried tf.functions, but the graph I got is too fine-grained, as many ops (Adam for example) are splited into smaller operators (add, mul, div etc.). So is there any methods that I can get a coarse-grained op-level graph? Thanks a lot!
Asked
Active
Viewed 104 times
0
-
You can use conceptual graph using Tensorboard, for more details you can visit [https://www.tensorflow.org/tensorboard/graphs](https://www.tensorflow.org/tensorboard/graphs) – Aug 18 '21 at 08:16
-
@Tfer3 You mean I can extract the graph info from tensorboard event file? I have noticed the tensorboard conceptual graph, but I am not sure which APIs the tensorboard event file provide. – Jason Aug 19 '21 at 09:16