lhnguyen102 / cuTAGI

CUDA implementation of Tractable Approximate Gaussian Inference
MIT License
24 stars 9 forks source link

Transitioning changes from V0 to V1 for heteroscedastic cases only #61

Closed bhargobdeka closed 3 months ago

bhargobdeka commented 3 months ago

changes are made to add “out_gain” in the files below wherever noise_gain was also found. net_init.cpp net_prop.cpp struct_var.h pytagi/tagi_network.py python_api_cpu.cpp python_api.cu

These changes should only affect heteroscedastic case study as the changes were made in the “gaussian_param_init_ni” function inside net_prop.cpp as shown in the below screenshot.

Screen Shot 2024-04-08 at 6 04 52 PM

Please review these changes and we can proceed for merging into the main branch.