-
Notifications
You must be signed in to change notification settings - Fork 235
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Torch FX] Add Special FX Metatype in Torch Operator Metatypes #2976
base: develop
Are you sure you want to change the base?
Conversation
@@ -120,15 +135,15 @@ def create_nncf_graph(model: torch.fx.GraphModule) -> PTNNCFGraph: | |||
|
|||
for source_node in model.graph.nodes: | |||
node_type, node_metatype = GraphConverter._get_node_type_and_metatype(source_node, model) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Could you suggest how to take function namespace into account when defining metatype?
cc' @daniil-lyakhov @AlexanderDokuchaev
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Alright, I will check for possible ways to map with function namespaces as well.
Apparently covered by #2891 |
Changes
FXEmbeddingMetatype
for torch FX embedding metatypeReason for changes
The current torch embedding metatype, which maps to
torch.nn.functional.embedding
has weight node on port 1, but when usingtorch.nn.Embedding
with torch fx, since it has weight on port 0, the metatype should be defined accordingly.