Skip to content

optimized image_to_source with numpy vectorization, Bug Fix and GPU grid caching#161

Open
kireetibabu07 wants to merge 2 commits intoML4SCI:mainfrom
kireetibabu07:main
Open

optimized image_to_source with numpy vectorization, Bug Fix and GPU grid caching#161
kireetibabu07 wants to merge 2 commits intoML4SCI:mainfrom
kireetibabu07:main

Conversation

@kireetibabu07
Copy link

@kireetibabu07 kireetibabu07 commented Feb 28, 2026

Hello sir,
while exploring the components, I noticed and made some required changes to optimize that.

Changes made:

  1. Vectorization: Replaced nested loops in Physics.py with NumPy vectorization for faster image reconstruction (in physics.py file)
  2. Memory Optimization: Added coordinate grid caching in PhysicsInformedEncoder.py to reduce redundant GPU allocations (in PhysicsInformedEncoder.py file)
  3. Bug Fix: Resolved NameError by adding missing imports for Physics and TransformerLSABlock (in PhysicsInformedEncoder.py file)

If possible i would love to hear feedback from you.

@kireetibabu07 kireetibabu07 reopened this Feb 28, 2026
@kireetibabu07 kireetibabu07 changed the title optimized image_to_source with numpy vectorization optimized image_to_source with numpy vectorization and GPU grid caching Feb 28, 2026
@kireetibabu07 kireetibabu07 changed the title optimized image_to_source with numpy vectorization and GPU grid caching optimized image_to_source with numpy vectorization, Bug Fix and GPU grid caching Feb 28, 2026
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.

1 participant