Running into problems during DataflowPartitioning #1051
Unanswered
Adityasrinivas24
asked this question in
Q&A
Replies: 1 comment 1 reply
-
I have never dealt with a model with Maxpool. However, the transpose nodes should be absorbed into the multithreshold. Also, the repeated Multithreshold nodes can be collapsed into a single node. Check this link and refer the transformations there it might help you. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi, I'm following the cnv_end2end example for my custom defined Neural Network. I streamlined the model as shown in the example and tried using
CreateDataflowPartition()
but ran intoAssertionError: cycle-free graph violated: partition depends on itself
.I went through a couple of older issues but they didn't fix my problem. Is there a standard flow that i can use to perform the transformations? How do i figure out problems in netron or other utility?
Ive also attached the network visualised in netron after streamlining. Any help will be appreciated .
Thank you
Beta Was this translation helpful? Give feedback.
All reactions