# CATCH

**Repository Path**: Tomhappy/CATCH

## Basic Information

- **Project Name**: CATCH
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-03-15
- **Last Updated**: 2026-03-15

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# CATCH
The code for "CATCH: Causal Attention Enhanced Robust Hyperbolic Heterogeneous Graph Representation Learning" (under review)
## Requirements
This repository has been tested with the following packages:
+ Python == 3.7.13
+ PyTorch == 1.12.1+cu113
+ DGL == 0.9.1
## Important Hyper-parameters
+ `curvature`: the curvature of the hyperbolic space.
+ `causality_lambda`: the balancing coefficient of the causality-related term.
+ `lr`:the learning rate.
+ `weight_decay`: the value of weight-decay.
+ `aggre_type`: the type of the information aggregation.
## How to run
For example:
run `multi.py`