Skip to content

Conversation

@chilo-ms
Copy link
Contributor

@chilo-ms chilo-ms commented Jan 26, 2026

This PR makes following changes:

  • Complete GetSupportedList() in GetCapability() so that it determines whether the entire graph or a subgraph (i.e., a subset of nodes from the original graph forming a new graph) should be assigned to the plugin TensorRT EP, based on the results from calling the TensorRT parser.
  • Before calling "OrtGraph to ModelProto" utility function, do the topological sort of the nodes first and then passing the "newly sorted" graph to the utility function.
  • Move source code under src/ folder

size_t max_workspace_size_ = 1 << 30; // 1GB
bool fp16_enable_ = false;
bool int8_enable_ = false;
bool bf16_enable_ = false;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is bf16 support fully implemented? it looks like this is only used to determine whether to set the flag nvinfer1::NetworkDefinitionCreationFlag::kSTRONGLY_TYPED.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch, added the missing parts for bf16 support.

@chilo-ms chilo-ms merged commit b201380 into main Jan 27, 2026
30 of 32 checks passed
@chilo-ms chilo-ms deleted the chi/complete_get_capability_for_plugin_trt_ep branch January 27, 2026 23:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants