Flow Matching

Creator
Creator
Seonglae Cho
Created
Created
2024 Dec 28 14:28
Editor
Edited
Edited
2024 Dec 28 16:57
Refs
Presents a method to train
CNF
without simulation, training the model by regressing the
Vector Field
of fixed conditional probability paths. The path is generally expressed as a probability flow (
Vector Field
) that varies with time tt.
This approach improves upon the sampling efficiency issues present in existing diffusion models and enables a more efficient generation process by utilizing diverse probability paths.
Flow Matchings
 
 
 
 
 
 

Recommendations