> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/sdgm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/sdgm/_mcp/server.

# positional_encodings

#### `positional_encodings: <list[list[hop | random_gnn]]>` (Optional)

#### Description

Determines whether positional encodings are used in the GraphTransformer model.

* `hop`: Adds a positional embedding to each node depending on the hop distance to the target entity.
* `random_gnn`: Computes positional embeddings with a Positional GNN based on initial random embeddings and adds them to each node.

#### Supported Task Types

* Node classification and regression

#### Default Values

| [run\_mode](/run-mode) | Default Value         |
| :--------------------- | :-------------------- |
| FAST                   | `[[hop, random_gnn]]` |
| NORMAL                 | `[[hop, random_gnn]]` |
| BEST                   | `[[hop, random_gnn]]` |