ONNX-Tensorflow Frontend Tutorial#27
Conversation
|
|
| " - *Graph definition*: You need to obtain information about the graph definition in the form of GraphProto. The easiest way to achieve this is to use the following snippet of code as shown in the example training script:\n", | ||
| "```\n", | ||
| " with open(\"graph.proto\", \"wb\") as file:\n", | ||
| " graph = tf.get_default_graph().as_graph_def(add_shapes=True)\n", |
There was a problem hiding this comment.
For some reason this indentation isn't showing up when this is rendered as markdown
|
@jamesr66a seems to be fine now. |
|
cc @bddppq Also, did we reach conclusion which of the TF frontends we advertise here (this vs https://github.com/onnx/tensorflow-onnx ) cc @prasanthpul |
|
@dzhulgakov not really but I don't see any problem advertising both for now and resolve this issue during our next meeting? |
|
I guess it's fine. I'm just making sure there's some longer term plan here. Thanks for adding the example! |
|
@tjingrant Hope you got the email I sent you last week to connect and discuss? |
|
@prasanthpul ahh it was you! Sorry I'm having trouble mapping your github handle to your email address; also that's my personal email, so sorry for not attending it as much as I should be. |
|
@jamesr66a ? Shall we move forward? |
|
@jamesr66a @prasanthpul can you merge this? It's been a while now. Do we have any additional concerns? |
5 participants
This is a tutorial for the frontend part of onnx-tensorflow.
cc @fumihwh, @jamesr66a