policy

Aesthetic-Crop-Selection-with-RLHF

alikasiri1 · PyTorch

or hover any field below to flag it

Overview

Name
Aesthetic-Crop-Selection-with-RLHF
Author
alikasiri1
Framework
PyTorch
License
unknown
Skill type
navigation
Evidence level
untested
Task description
explores automatic image cropping and composition by combining a pre-trained NIMA model with reinforcement learning from human feedback (RLHF). The system uses NIMA to score candidate crops and then refines its selection policy through human-aligned reinforcement learning, aiming to produce visually

Spaces

Action space
other · 0-dim · 0Hz
Observation space
  • type: other

Links

HuggingFace repo
null
Paper (arXiv)
null

Compatible robots

20

Compatible environments

0

No environments list Aesthetic-Crop-Selection-with-RLHF yet.

Datasets that reference this policy

0

No datasets reference Aesthetic-Crop-Selection-with-RLHF yet.